july 2020 bulletin · bank of greece 21, e. venizelos avenue gr-102 50 athens economic analysis...
TRANSCRIPT
BA
NK
OF
GR
EEC
E
EUR
OSY
STE
MJU
LY
20
20
EC
ON
OM
IC B
UL
LE
TIN
No
51
JULY2020
ECONOMICBULLETIN
No 51
JULY2020
ECONOMICBULLETIN
No 51
BANK OF GREECE21, E. Venizelos AvenueGR-102 50 Athens
www.bankofgreece.gr
Economic Analysis and Research Department - SecretariatTel. +30 210 320 2393Fax +30 210 323 3025
Printed in Athens, Greeceat the Bank of Greece Printing Works
ISSN 1105 - 9729 (print)ISSN 2654 - 1904 (online)
CONT ENT S COVID-19 AND OTHER PANDEMICS: A LITERATURE REVIEW FOR ECONOMISTS 7Sofia Anyfantaki Hiona BalfoussiaDimitra DimitropoulouHeather GibsonDimitris PapageorgiouFilippos PetroulakisAnastasia TheofilakouMelina Vasardani
SOVEREIGN CREDIT RATINGS AND THE FUNDAMENTALS OF THE GREEK ECONOMY 43Dimitris MalliaropulosPetros Migiakis
THE CURRENT ACCOUNT ADJUSTMENT IN GREECE DURING THE CRISIS: CYCLICAL OR STRUCTURAL? 73Constantina BackinezosStelios PanagiotouChristos Papazoglou
D-EURO: ISSUING THE DIGITAL TRUST 91Yorgos Korfiatis
WORKING PAPERS(January – July 2020) 127
ARTICLES PUBLISHED IN PREVIOUS ISSUES OF THE ECONOMIC BULLETIN 131
51Economic BulletinJuly 20206
Sofia Anyfantaki
Hiona Balfoussia
Dimitra Dimitropoulou
Heather Gibson
Dimitris Papageorgiou
Filippos Petroulakis
Anastasia Theofilakou
Melina Vasardani
Economic Analysis and Research Department
ABSTRACTWe provide a comprehensive review of the literature on the economic impact of pandemics andidentify the transmission channels at play. The primary channel comes from the supply side, aspandemics reduce both the quantity and the quality of labour. They can also lead to a destruc-tion of capital, as businesses close and investment is curtailed. On the demand side, consump-tion is particularly vulnerable to the impact of both reduced income and declining consumer con-fidence. A third channel works through the financial system. While the natural rate of interestmight be expected to fall, leading to a period of low interest rates, financial institutions are likelyto come under stress. Rising uncertainty, along with an increase in the number of borrowers withdebt servicing difficulties, may dampen investment and generate a liquidity squeeze, exacerbatingthe demand effects of the pandemic. All three channels work to reduce current and potential out-put. Spillovers and asymmetries can explain the varying impact of the pandemic across countries,but it seems that open economies, embedded in global value chains, are especially vulnerable.Nonetheless, the literature provides ample evidence on how to limit the impact of pandemics usingmonetary and fiscal policy combined with measures to ease liquidity constraints on the financialsector. In the context of the EU, the coordination and mutualisation of the policy response tothe pandemic can prove to be very beneficial.
Keywords: COVID-19; pandemic; lockdown; social distancing; propagation; spillovers
JEL classification: E2; E3; E5; G1
51Economic Bulletin
July 2020 7
COV I D - 19 AND O THER P ANDEM I C S : A L I T E R A TUR E R E V I EW FOR E CONOM I S T S
Σοφία Ανυφαντάκη
Χιόνα Μπαλφούσια
Δήμητρα Δημητροπούλου
Heather Gibson
Δημήτρης Παπαγεωργίου
Φίλιππος Πετρουλάκης
Αναστασία Θεοφιλάκου
Μελίνα Βασαρδάνη
Διεύθυνση Οικονομικής Ανάλυσης και Μελετών
ΠΕΡΙΛΗΨΗΗ μελέτη αποτελεί μια επισκόπηση της βιβλιογραφίας όσον αφορά τις οικονομικές επιπτώσειςμιας πανδημίας και των διαύλων μετάδοσής τους. Ο κύριος δίαυλος προέρχεται από την πλευράτης προσφοράς, καθώς οι πανδημίες μειώνουν τόσο την ποσότητα όσο και την ποιότητα της προ-σφερόμενης εργασίας. Ενδέχεται επίσης να οδηγήσουν σε καταστροφή κεφαλαίου, καθώςκάποιες επιχειρήσεις αναγκάζονται να μειώσουν τις επενδύσεις τους ή και να παύσουν τη λει-τουργία τους. Από τη σκοπιά της ζήτησης, είναι ιδιαιτέρως πιθανόν να πληγεί η κατανάλωση,λόγω της μείωσης του διαθέσιμου εισοδήματος και της καταναλωτικής εμπιστοσύνης. Ένας τρί-τος δίαυλος είναι χρηματοπιστωτικός. Παρότι το φυσικό επιτόκιο είναι πιθανόν να μειωθεί, προ-μηνύοντας μια περίοδο χαμηλών επιτοκίων, είναι αναμενόμενο ότι τα χρηματοπιστωτικά ιδρύ-ματα θα υποστούν πιέσεις. Η αύξηση της αβεβαιότητας και η άνοδος του ποσοστού των δανει-οληπτών σε δυσχέρεια οδηγούν σε περιορισμό της ρευστότητας και επιτείνουν τις αρνητικές επι-πτώσεις της πανδημίας στη ζήτηση. Διαμέσου των τριών αυτών διαύλων, τόσο το παραγόμενοόσο και το δυνητικό προϊόν υφίστανται μείωση. Οι οικονομικές επιπτώσεις μιας πανδημίας ενδέ-χεται να διαφέρουν από χώρα σε χώρα, λόγω ασυμμετριών και φαινομένων διάχυσης. Ωστόσο,φαίνεται ότι οι ανοικτές οικονομίες, που είναι ενταγμένες στις διεθνείς αλυσίδες αξίας, είναιιδιαιτέρως ευάλωτες σε αυτές τις επιδράσεις. Στο πλαίσιο της Ευρωπαϊκής Ένωσης, ο συντο-νισμός και η αμοιβαιοποίηση των οικονομικών μέτρων για την αντιμετώπιση των επιπτώσεωντης πανδημίας μπορεί να αποβεί εξαιρετικά επωφελής.
51Economic BulletinJuly 20208
COV I D - 19 ΚΑ Ι ΑΛΛΕΣ ΠΑΝΔΗΜ ΙΕΣ : ΕΠ Ι ΣΚΟΠΗΣΗΤΗΣ Β Ι ΒΛ ΙΟ ΓΡΑΦ ΙΑΣ Γ Ι Α Ο ΙΚΟΝΟΜΟΛΟΓΟΥΣ
1 INTRODUCTION
The COVID-19 pandemic is arguably thelargest peace-time potential threat to life on aglobal scale for a century. In probably the mostwell-known simulation study, Ferguson et al.(2020) argue that COVID-19 could be thedeadliest pandemic since the Spanish Flu, withhalf a million deaths in the UK and over 2 mil-lion in the US. COVID-19 and the unprece-dented non-pharmaceutical interventions(NPIs) and preventative policies to contain itplace urgency on trying to gauge the likely eco-nomic impacts.1
While fatality rates are difficult to gauge, dueto large uncertainty about the number of cases,it seems that COVID-19 is unlikely to be asdeadly as the Spanish Flu, which claimed atleast 50 million lives (see Table 1). A notabledifference is that the Spanish Flu primarilyaffected prime-age workers, which suggestsmore severe economic impacts of the 1918influenza pandemic, particularly for potentialoutput.2 But 21st-century epidemics can spreadmore widely and more quickly, having aruinous impact on the economy of the affectedcountry and a contagion effect on the globaleconomy. The more complex nature of modernglobal supply chains, the intense mobility of
human populations, the greater role of serv-ices, and the improvements in information andcommunication technologies are importantfactors for understanding the macroeconomiceffects of COVID-19. While 102 years after theSpanish Flu, there is a more sound knowledgeof infectious diseases as well as a wider rangeof potential interventions effective in pre-venting their spread, most of the countriesworldwide still face the same challenge of mit-igating the disruptive impacts of the pandemic.
A large set of papers has emerged on macro-economic issues surrounding the COVID-19pandemic. The world has entered intouncharted territory, with little history to guidepolicy makers on what the expected economicfallout will be, which societal interventions arewarranted to contain its spread, and how a sys-
51Economic Bulletin
July 2020 9
COV I D - 19 AND O THER P ANDEM I C S : A L I T E R A TUR E R E V I EW FOR E CONOM I S T S *
Sofia Anyfantaki
Hiona Balfoussia
Dimitra Dimitropoulou
Heather Gibson
Dimitris Papageorgiou
Filippos Petroulakis
Anastasia Theofilakou
Melina Vasardani
Economic Analysis and Research Department
* We would like to thank Dimitris Malliaropulos for useful commentson earlier drafts as well as our colleagues in the Economic Analysisand Research Department for their suggestions when we presentedthe paper. The present study draws on research published up untilend-April 2020. The views expressed are those of the authors anddo not necessarily reflect those of the Bank of Greece. The authorsare responsible for any errors or omissions.
1 The 20th century has witnessed two influenza pandemics since theSpanish Flu of 1918: the Asian flu of 1957 and the Hong Kong fluof 1968. The 21st century has seen a number of pandemics, mostnotably the Severe Acute Respiratory Syndrome (SARS) in 2002,H1N1 (“bird flu”) in 2009, the Middle East Respiratory Syndrome(MERS) in 2012, and Ebola, which peaked in 2014-16.
2 99% of the victims of the Spanish Flu were below 65 years of age,and around half of them were aged between 20 and 40.
temic response should be organised. What isthe economic impact of a pandemic? Are theeconomic effects temporary or persistent?What is the impact of public health responseson the economy? Does the early and extensiveuse of NPIs, such as social distancing and lock-downs, which slow the spread of the pandemic,reduce its medium-term economic severitydespite the inevitably heavy short-run toll?3
There are several policy proposals, with a largenumber of them collected in Baldwin and diMauro (2020a). For example, Gourinchas(2020) argues that “flattening the infectioncurve inevitably steepens the macroeconomicrecession curve”. Although the measures thathelp address the health crisis can make theeconomic crisis worse ―at least in the shortrun― the consensus amongst economistsseems to be that containment is the appropri-
ate policy. Economic policy can act decisivelyto limit the economic damage and thus “flat-ten the curve”.
The remainder of this paper is structured asfollows: in Section 2 we seek to shed some lighton the above questions by, first, focusing onthe literature that looks at the economicimpact of past pandemics, including the BlackDeath and the Spanish Flu. We then move onto Section 3 to explore how economists aremodelling the impact of the current pandemicwithin their empirical models. An examinationof both strands of the literature allows us toidentify the channels through which pandemics
51Economic BulletinJuly 202010
Black Death 1347 1352 75,000,000
Plague in Spain 1596 1631 600,000-700,000
Italian Plague 1629 1631 280,000
Great Plague of Sevilla 1647 1652 2,000,000
Naples Plague 1656 1656 1,250,000
Great Plague of London 1665 1666 100,000
Great Northern War Plague 1700 1721 176,000-208,000
Great Plague of Marseille 1720 1722 100,000
First Asia Europe Cholera Pandemic 1816 1826 100,000
Second Asia Europe Cholera Pandemic 1829 1851 100,000
Russia Cholera Pandemic 1852 1860 1,000,000
Fourth Cholera Pandemic 1863 1875 600,000
Global Flu Pandemic 1889 1890 1,000,000
Sixth Cholera Pandemic 1899 1923 80,000
Encephalitis Lethargica Pandemic 1915 1926 1,500,000
Spanish Flu 1918 1920 100,000,000
Asian Flu 1957 1958 2,000,000
Hong Kong Flu 1968 1969 1,000,000
H1N1 Pandemic 2009 2009 203,000
Event Start End Death toll
Table 1 Nineteen deadly pandemics since early modern times
Source: Cirillo and Taleb (2020). Note: The list is not exhaustive, but focuses on pandemics on which data are available.
3 NPIs intend to reduce infectious contacts between persons andform an integral part of plans to mitigate the impact of an influenzapandemic. The potential benefits of NPIs are supported by bothmathematical models and historical evidence on the impact of suchinterventions in past pandemics (see for example Bootsma andFerguson 2007; Markel et al. 2008; Hatchett et al. 2007).
impact economies. In particular, it is possibleto quantify the impact on various macroeco-nomic aggregates, including GDP, inflation,wages, poverty, trade, non-performing loans(NPLs) and social capital. The literature alsopoints to potential policy measures that canmitigate the impact of the pandemic. Section4 draws on the assessment recently made byinternational institutions to discuss the role ofasymmetries and the spillover effects ofCOVID-19. Section 5 draws on all previoussections to summarise the key transmissionchannels and provide economic policy impli-cations. Section 6 concludes.
2 REVIEW OF RESEARCH ON THE ECONOMICIMPACT OF PAST PANDEMICS
2.1 EMPIRICAL RESEARCH ON PAST PANDEMICS
The economic history literature on the rela-tionship between pandemics and economicoutcomes is hardly new (see Tables 2 and 3 fora summary). Most historical studies have typ-ically focused on one event in one country orregion and have traced local outcomes for upto a decade at most. Due to the absence ofrecent pandemics, work has primarily relied onaggregated data at the regional or nationallevel.4
By far the most severe pandemic in terms offatality rates was the Black Death (Plague),which decimated anywhere between 30% and60% of Europe’s population in the 14th cen-tury.5 Most work has shown that, by sharplyreducing the size of the working population,the Plague led to a substantial increase in nom-inal wages for workers that persisted into the15th century. Real wages took considerablylonger to reach pre-Plague levels, especially forskilled workers (Munro 2005), as the supplyshock-driven food inflation was higher thanwage inflation. One reason for this was insti-tutional: governments in several regions insti-tuted formal wage restraints, particularly forrural workers (Munro 2005). The impact onper capita income is less clear, since rents also
fell in the aftermath of the Black Death (Hir-shleifer 1987; see also Robbins 1928).6 Jedwabet al. (2019), using a granular dataset for alarge sample of European cities, argue thatpopulation changes fit well a Malthusiangrowth paradigm, where production relies onfixed factors (land and natural resources) withlittle to no technological improvements, mak-ing the lessons learnt from the Black Death oflittle value for modern pandemics.
Due to its unprecedented devastation, theBlack Death is likely to have for ever alteredthe economic and social landscape of Europe.Jedwab et al. (2019) provide evidence that ruralvillages were abandoned, as inhabitantsmoved to more affected cities to exploit theabundance of fixed factors of production.Voigtländer and Voth (2012) provide the mostshocking example of persistent social effects.Scapegoating of Jews during the Black Deathwas quite prevalent in Central Europe, andmass killings took place in several cities in Ger-many. The authors show that cities with anti-Jewish pogroms during the Black Death hadmarkedly higher patterns of violent anti-semitism in the 1920s. While this is not causalevidence of Black Death affecting antisemitism,it is strong evidence of the endurance of cul-tural traits, which are likely to have been insome way shaped by the Black Death.
The most widely cited major pandemic was theSpanish Flu, due to its relatively recent occur-
51Economic Bulletin
July 2020 11
4 The economics pandemic literature has typically not consideredHIV/AIDS together with other major pandemics, despite its large-scale death toll (over 30 million). The economic effects ofHIV/AIDS are likely very different because it is much more diffi-cult to transmit than pandemics caused by flu, cholera or plague,and can hence be more manageable; flu episodes can occur verysuddenly, giving rise to outbreaks. HIV/AIDS is also slow-acting(death occurs within years versus days for the rest). The WHO hasclassified HIV/AIDS as a global epidemic, instead of a pandemic.
5 The Black Death ―a combination of bubonic and pneumonicplagues― killed roughly one-quarter of the Western Europeanpopulation between 1348 and 1351, and recurring epidemicscontinued to inflict high death tolls on the continent over the nextquarter-century. It is unclear whether the Black Death caused moreabsolute fatalities than the Spanish Flu, but the relative magnitudein terms of total population was far higher.
6 Bloom and Mahal (1997) re-examine the effect of the Plague onthe wages of unskilled agricultural workers in England duringepidemics that occurred between 1310 and 1449. They find apositive but statistically insignificant relationship between realwages and population growth, with similar results for France.However, their study is hampered by its very small sample size.
51Economic BulletinJuly 202012
Black Death
Munro (2005)
Real wages initially fell buteventually grew substantiallyand persistently.
Negative effect on realwages despite highernominal wages.
Large gains in realwages lasting for a cen-tury.
Labour shortages led to poor harvests in theshort run, raising food inflation faster thanwages. Wage controls imposed by landowners.In the long run, these effects waned and theshortage of labour relative to capital led topersistent wage gains.
Voigtländer and Voth (2012)
Antisemitism in the 1920s higherin cities with anti-Jewish pogromsin Black Death.Cities with high levels of trade orimmigration show less persistence.
Local continuity of vio-lence against Jews over600 years (votes for theNazi Party, deportationsafter 1933, attacks onsynagogues).
Persistence of cultural traits.
Jedwab et al. (2019)
Affected cities recovered theirpopulation over the very long run.
Negative effects on citysize.
Recovery according toMalthusian model.
Migration to areas with labour shortages andabundant fixed production factors (tradepotential).
Spanish Flu
Brainerd and Siegler (2003)
Substantial macroeconomiceffects in the US, even after con-trolling for differences acrossstates. Likely a contributing fac-tor to post-WWI recessions.
Substantial business fail-ures which caused theeconomy to be belowtrend, on average,between 1919 and 1921.
One more deathresulted in an averageannual increase of atleast 0.2% in growthover the next 10 years.
Effect of prime-age influenza mortality rates(rather than overall mortality rates) on growth.Higher capital deepening (business failures’effect smaller than labour supply effect), lowerlabour force growth, increased investment inhuman capital.
Almond (2006)
Children of infected mothersfared worse in later life.
Depressed human capi-tal.
Sickness during pregnancy.
Garrett (2009) Wages in the US grew in relativeterms in more exposed areas.
Approximately 4% of total wage growth(1914-19) is attributed toinfluenza mortalities.
Reduction in labour supply raised marginalproduct of labour.However, it is not always clear to what extentthe results are attributable to WWI.
Karlsson et al. (2014)
Sweden; persistent decline inrental income and increase inpoverty; no effect on earnings.More affected regions grewslower after the pandemic.
Rental income decline. Rental income declineand higher povertyrates.
Labour quality fell in affected areas, mitigat-ing wage growth due to labour scarcity. Alsoexplains higher poverty and lower marginalproduct of capital.
Correia et al. (2020)
Areas that were more severelyaffected see a sharp and persist-ent decline in real economicactivity, controlling for otherfactors.Early and extensive NPIs haveno adverse effect on local out-comes; instead, a relativeincrease in real economic activ-ity after the pandemic.
18% decline in state man-ufacturing output; rise inbank charge-offs. Reacting 10 days earlierincreases employment by5% in post-period; 50additional days raiseemployment by 6.5%; ones.d. higher number of daysinduces a 7.5% larger localbanking sector after 1918.
More affected areasremain depressed rela-tive to less exposedareas from 1919through 1923.The reduction in bankassets is persistent.
Reductions in both supply and demand.NPIs can have economic merits, beyond low-ering mortality.Important channel of transmission of bothdemand and supply shocks could have beenthe banking sector.
Barro (2020)
NPIs have no significant impacton cumulative mortality in the US.
Short duration; but peakmortality does fall.
Velde(2020)
Little short-term economiceffects of the pandemic in theUS. Recessions short and mod-est.
Considers a number of high-frequency indica-tors to show little impact during the pan-demic, but large recession after the pandemic.Puzzling results given other papers showinglarge long-term effects.
Barro et al. (2020)
Declines in GDP and consump-tion. Effects are fully permanentor fully temporary or somewherein between. Decreased realised real returnson stocks and, especially, onshort-term government bills. Higher inflation at least tem-porarily.
Increased inflation rates(at least temporarily).Effect on stock negativebut not significant; nega-tive and significant effecton bonds.
Reduction of real percapita GDP by 6%.Larger effects for con-sumption. For the USonly 1.5 % decrease inGDP and 2.1% in con-sumption.Effects on asset returnnot reversing
Effects of the Great Influenza Pandemic andWWI on economic growth (treated as (mostly)exogenous variables), gauged by growth rates ofreal per capita GDP and real per capita con-sumption (personal consumer expenditure). WWI and Great Influenza Pandemic are viewedas unanticipated and contemporaneously per-ceived as having some persistence but ultimatelybeing temporary.
Le Moglie et al.(2020)
Spanish Flu mortality associatedwith lower social trust.
Long-run impact ondescendants of immi-grants.
Social distancing measures impeded socialinteractions. Persistence of cultural traits.
All pandemics
Jordà et al. (2020)
Large and persistent reductionin natural interest rate; oppositefor real wages.
Limited effects in theshort run (until tenyears).
Substantial effects last-ing for at least fourdecades, longer insome countries.
Excess supply of capital relative to labour;wars (which destroy capital) opposite effect.With sufficiently low depreciation, highergrowth potential can be accommodated withlow investment, leading to lower natural rates.Wage effects as in Munro (2005).
Summary of findings Short-term impact Long-term impact Channel of transmission
Table 2 Summary of main empirical papers on past pandemics
rence and truly global nature.7 Brainerd andSiegler (2003) were the first to examine itseffects8 on subsequent growth. Using data froma sample of US states, they find a positive cor-relation, even after controlling for a number ofdifferences across states.9 They suggest thatone more death per thousand resulted in anaverage annual increase of at least 0.2 per-centage point in the rate of economic growthover the next ten years. However, they find thatflu deaths in 1918 and 1919 among prime-ageadults are a significant predictor of businessfailures in 1919 and 1920, implying that theeconomy may have been below trend, on aver-age, between 1919 and 1921. In other words,some of the growth from 1919-21 to 1930 isonly a return to trend after this large tempo-rary shock. The concurrent presence of higherbusiness failures immediately after the pan-demic and higher subsequent growth inaffected areas may reflect a combination offactors: higher capital deepening (if thedestruction of capital as a result of businessfailures were small enough, relative to thereduction in labour supply); lower labour forcegrowth (as the young were especially affected);increased investment in human capital; or sim-ple convergence. At the same time, long-termeffects are hard to infer due to the boom of the1920s and the subsequent crash of financialmarkets in 1929.
Garrett (2009) examines the immediate effectof influenza mortalities on manufacturingwages in US cities and states, jointly with theeffect of World War I (WWI). The hypothe-sis is that influenza mortalities, by reducingthe supply of manufacturing workers, raisedthe marginal product of labour and thus realwages. Since, in the short term, labour immo-bility across cities and states is likely to haveprevented wage equalisation across states, asubstitution to capital is unlikely to haveoccurred. The study finds that states and citieswith greater mortalities experienced greaterwage growth – roughly 2 to 3 percentagepoints for a 10% change in per capita mor-talities. Approximately 4% of total wagegrowth from 1914 to 1919 is attributed to
influenza mortalities. However, it is notalways clear to what extent the results attrib-utable to the effect of influenza are distinctfrom the impact of WWI.
More recently, Karlsson et al. (2014), using adifference-in-differences analysis of high-qual-ity administrative data from Sweden, estimatethe effects of the 1918 influenza pandemic onearnings, capital returns and poverty. They findthat the pandemic led to a significant increasein poverty rates, and a reduction in capitalreturns; but, contrary to others, they do notfind significant effects on earnings. At thesame time, they show that more affectedregions grew slower in the aftermath of thepandemic. Thus, the combination of fallingcapital income and growth in more affectedareas but with no difference in earnings mayexplain the labour supply reaction, i.e. thereduction in average worker quality. In thisway, the study shows that labour heterogene-ity needs to be taken into account whenanalysing the effects of a pandemic.
A few recent papers have revisited the effectof the Spanish Flu on the US economy, withmixed results. Correia et al. (2020) study a vari-ety of economic outcomes using city-level vari-ation in mortality. They find that more exposedareas experienced a sharper and more per-sistent decline in economic activity relative toother areas, controlling for possible contem-poraneous shocks, interacted with local char-acteristics. Consistent with Brainerd andSiegler (2003), they find that severely and mod-erately affected areas had similar levels of pop-ulation, employment, and income per capitabefore 1918. They also address endogeneityconcerns by exploiting the fact that regions dif-fered in susceptibility to influenza outbreaks,
51Economic Bulletin
July 2020 13
7 Why the wave was so deadly ―with mortality rates 5 to 20 timeshigher than normal― and why it primarily affected young adultsis still unclear, despite much recent research on the 1918 influenzaepidemic by microbiologists.
8 Bloom and Mahal (1997) examine the effect of the 1918 influenzapandemic in India, which experienced an estimated 17 to 18 milliondeaths, and find no relationship between the magnitude of thepopulation decline and changes in acreage sown per capita.
9 Brainerd and Siegler (2003, p. 7) conclude that “the statisticalevidence also supports the notion of influenza mortality as anexogenous shock to the population”.
51Economic BulletinJuly 202014
Mel
tzer
et
al.
(199
9)
Gro
ss a
ttac
k ra
te 1
5%-3
5%.
Pre
defi
ned
prob
abili
ty d
istr
ibut
ions
for
a se
t of i
nput
var
iabl
es b
y ag
ean
d ri
sk g
roup
.
Fou
r op
tion
s of
vac
cina
tion
sche
mes
.
Mon
te C
arlo
sim
ula-
tion
.
Eco
nom
ic im
pact
of
vacc
ine-
base
d in
ter-
vent
ions
.
US
USD
71.
3-16
6.5
billi
on
Dis
ease
-ass
ocia
ted
med
ical
cos
ts, i
ndir
ect c
osts
from
loss
es o
f tim
ean
d in
com
e by
car
eers
, and
mor
bidi
ty/m
orta
lity
cost
s m
easu
red
asex
pect
ed fo
rego
ne li
feti
me
earn
ings
.
Exc
ludi
ng d
isru
ptio
ns in
com
mer
ce a
nd s
ocie
ty.
Los
s of
life
acc
ount
ed fo
r ap
prox
imat
ely
83%
of a
ll ec
onom
ic lo
sses
at a
ny g
iven
atta
ck r
ate.
The
larg
est r
etur
ns c
ome
from
inte
rven
tions
that
pre
vent
the
larg
est n
umbe
r of
dea
ths.
Vac
cina
ting
pri
orit
ies
shou
ld b
e se
tde
pend
ing
on th
e po
licy
obje
ctiv
e.
Oth
er m
ulti
plie
r ef
fect
s re
sult
ing
from
disr
upti
ons
in c
omm
erce
and
soc
iety
nee
dal
so to
be
valu
ed.
Lar
ge u
ncer
tain
ty a
bout
the
gros
s at
tack
rate
of a
n ac
tual
pan
dem
ic.
Lee
and
McK
ibbi
n(2
004)
Tem
pora
ry s
hock
(6
mon
ths)
:in
crea
se in
cou
ntry
ris
k pr
emiu
m,
drop
in d
eman
d fo
r se
rvic
es a
ndin
crea
se in
cos
ts in
the
serv
ices
sec
-to
r.
Cou
ntry
-spe
cifi
c in
dex
of “
glob
alex
posu
re to
SA
RS”
.
Per
sist
ent s
hock
s of
equ
al s
ize
fade
out e
quip
ropo
rtio
nate
ly o
ver
a 10
-ye
ar p
erio
d.
Rat
iona
l exp
ecta
tion
s an
d fo
rwar
d-lo
okin
g in
tert
empo
ral b
ehav
iour
.
G-C
ubed
(A
sia-
Pac
ific
) m
odel
.
Eco
nom
ic c
osts
of
SAR
S.
Glo
bal
Tem
porary sho
ck
Glo
bal G
DP
loss
clos
e to
USD
40
bill-
lion
in 2
003;
Hon
g K
ong
(2.6
3% o
fG
DP
) w
ith
the
serv
-ic
es s
ecto
r th
e la
rges
tco
ntri
buti
ng fa
ctor
;
Chi
na (
1.05
%)
wit
hef
fect
eve
nly
spre
adac
ross
fact
ors.
Persisten
t sh
ock
Glo
bal G
DP
loss
clo
se to
USD
54
billi
on in
2003
.
Chi
na, H
ong
Kon
g, M
alay
sia,
the
Phi
lippi
nes,
Sing
apor
e an
d T
aiw
an la
rger
loss
vs
OE
CD
and
othe
rs lo
wer
GD
P lo
ss.
Gre
ater
cap
ital
out
flow
from
aff
ecte
d co
un-
trie
s in
to th
e le
ast a
ffec
ted
coun
trie
s.
Net
cap
ital o
utflo
ws
from
Chi
na a
nd H
ong
Kon
gar
e es
timat
ed to
0.8
% a
nd 1
.4%
, res
pect
ivel
y.
Per
sist
ence
in r
ise
of r
isk
prem
ium
cau
ses
larg
e ca
pita
l out
flow
; sha
rp c
ontr
acti
on in
inve
stm
ent l
eads
to p
ersi
sten
t dec
line
in p
ro-
duct
ion
capa
city
.
Med
ical
exp
endi
ture
s an
d de
mog
raph
icco
nseq
uenc
es o
f SA
RS
are
insi
gnif
ican
t.
The
res
ults
ove
rall
poin
t tow
ards
the
argu
-m
ent t
hat i
n a
com
plex
inte
rrel
ated
wor
ld,
a di
seas
e ou
tbre
ak m
ight
hav
e a
huge
eco
-no
mic
impa
ct fo
r th
e gl
obal
eco
nom
y an
dno
t onl
y fo
r th
e af
fect
ed e
cono
mie
s.
The
rec
essi
on is
foun
d to
last
for
a nu
m-
ber
of y
ears
aft
erw
ard.
Blo
om e
tal
. (20
05)
Mild
pan
dem
ic (
atta
ck r
ate
20%
;ca
se fa
talit
y ra
te 0
.5%
; one
yea
r).
Scen
ario
1: p
sych
olog
ical
impa
ctaf
fect
s de
man
d fo
r tw
o qu
arte
rs.
Scen
ario
2: f
our
quar
ters
.
Exo
geno
us c
onsu
mpt
ion
shoc
k of
3% a
nd c
ontr
acti
on in
the
expo
rt o
fse
rvic
es; d
eman
d sh
ock
only
inA
sian
cou
ntri
es; 2
-wee
k pe
riod
of
labo
ur s
uppl
y sh
ock
(not
wor
king
due
to th
e di
seas
e).
Oxf
ord
Eco
nom
icF
orec
asti
ng (
OE
F)
glob
al m
odel
.
Shor
t-ru
n ec
onom
icim
pact
of a
hum
an-
to-h
uman
infl
uenz
a.
Asi
a
Scena
rio 1
Asi
a: 2
.3%
GD
Pde
man
d sh
ock;
0.3
%G
DP
sup
ply
shoc
k.
Scena
rio 2
Mor
e se
vere
.
Glo
bal G
DP
shr
ink-
age
0.6%
: wor
ldre
cess
ion.
Glo
bal
trad
e of
goo
ds a
ndse
rvic
es c
ontr
acti
onby
14%
.
Lon
g-term
impa
ct
Eve
n af
ter
5 ye
ars
inSc
enar
io 2
Asi
a’s
GD
P g
row
th w
ill b
elo
wer
by
3.6
perc
ent-
age
poin
ts.
Dem
and
and
supp
ly e
ffec
t.
Dir
ect:
Asi
an c
onsu
mer
s re
duce
act
ivit
y.
Indi
rect
: res
t of w
orld
red
uces
con
sum
p-ti
on, i
mpa
ctin
g tr
ade
and
inve
stm
ent.
Ope
n ec
onom
ies
mor
e vu
lner
able
,ex
port
ers
of s
ervi
ces
hard
-hit
.
Tim
ely
polic
y re
spon
ses c
an h
elp
prev
ent a
ndm
itiga
te th
e ec
onom
ic im
pact
of a
pan
dem
ic.
Coo
pera
tion
and
coo
rdin
atio
n am
ong
coun
trie
s.
Assum
ptions
Mod
el/Aim
Scop
eGDP im
pact
Com
men
ts
Table
3Su
mm
ary
of p
aper
s w
hich
mod
el p
ast
pan
dem
ics
51Economic Bulletin
July 2020 15
McK
ibbi
nan
dSi
dore
nko
(200
6)
Mild
: sim
ilar
to 1
968-
1969
Hon
g K
ong
flu. M
oder
ate:
sim
ilar
to 1
957
Asi
anflu
. Sev
ere:
sim
ilar
to 1
918-
1919
Spa
n-is
h flu
. Ultr
a: S
pani
sh F
lu b
ut w
ithhi
gher
mor
talit
y ra
tes
for
olde
r pe
ople
.
Epi
dem
ic s
hock
s: in
dice
s of
pos
sibi
lity
of e
ach
seve
rity
to o
ccur
, sic
knes
sin
dex
for
mor
bidi
ty r
ate.
Att
ack
rate
30%
. R
educ
tion
in c
on-
sum
ptio
n, in
crea
se in
cos
t of d
oing
busi
ness
, inc
reas
e in
cou
ntry
ris
k pr
e-m
ia.
Glo
bal e
xpos
ure
of th
e co
untr
y to
the
dise
ase.
Asi
a P
acifi
c G
-Cub
ed m
odel
(ext
ende
d fo
r th
eU
K).
Est
imat
e m
acro
cons
eque
nces
of
influ
enza
pan
-de
mic
.
Glo
bal
Mild scen
ario
1.4
mill
ion
deat
hsw
orld
wid
e an
d gl
obal
GD
P w
ould
fall
byap
prox
imat
ely
0.8%
,w
ith m
orta
lity
and
mor
bidi
ty s
hock
sbe
ing
the
mai
n dr
iver
sas
sum
ing
that
mon
e-ta
ry p
olic
y ca
n ef
fec-
tivel
y co
ntam
inat
ede
man
d ch
ange
s.
Mod
erate scen
ario
As
seve
rity
of t
he p
an-
dem
ic in
crea
ses,
the
impo
rtan
ce o
f cos
tin
crea
ses
rise
s, r
esul
t-in
g in
a la
rger
shr
ink-
age
of g
loba
l GD
P, b
utw
ith th
e ne
gativ
eef
fect
bei
ng la
rger
for
Asi
a an
d de
velo
ping
coun
trie
s.
Mor
e su
bsta
ntia
l GD
Plo
sses
(9.
3% fo
r H
ong
Kon
g an
d 7.
3% fo
r th
eP
hilip
pine
s).
Severe scena
rio
Con
trac
tion
in m
ost
affe
cted
eco
nom
ies
refle
cts
muc
h la
rger
shoc
ks a
nd la
rge
real
-lo
catio
n of
glo
bal c
api-
tal.
Wor
ld is
sub
stan
tially
affe
cted
(up
to 2
6%co
ntra
ctio
n in
Asi
a).
Ultra scena
rio
Ove
r 14
2.2
mill
ion
deat
hs, a
nd in
com
e lo
sses
of o
ver
12%
of G
DP
(U
SD 4
.4 tr
illio
n) w
orld
wid
e.
Ove
r 50
% in
som
e de
velo
ping
cou
ntri
es s
uch
asH
ong
Kon
g in
200
6.
Lar
ge-s
cale
col
laps
e of
Asi
a ca
uses
glo
bal t
rade
flow
s to
dry
up
and
capi
tal t
o fl
ow to
saf
e ha
vens
.
Pri
ces
rise
in th
e sh
ort r
un, t
he r
esul
t dep
ends
on
whe
ther
dem
and
or s
uppl
y ef
fect
is la
rger
.
Mon
etar
y tig
htne
ss (
as a
res
pons
e to
dec
linin
g ou
tput
,in
flatio
n ch
ange
s or
exc
hang
e ra
te c
hang
es)
resu
lts in
high
er e
cono
mic
impa
ct a
nd m
ay h
ave
a gr
eat i
mpo
r-ta
nce
for
bond
mar
kets
toge
ther
with
fisc
al r
espo
nse.
Pol
icy
mak
ers
shou
ld in
vest
a lo
t in
aver
ting
an
out-
brea
k be
caus
e of
the
sign
ific
ant p
oten
tial
con
se-
quen
ces
for
the
glob
al e
cono
my.
Bur
ns e
t al.
(200
6)R
eplic
atio
n of
the
resu
lts o
f McK
ibbi
nan
d Si
dore
nko
(200
6).
Hum
an-t
o-hu
man
pan
dem
ic w
ith s
imi-
lar
mor
talit
y ra
te to
the
Span
ish
Flu
(1.0
8% o
f peo
ple
die
acro
ss th
e w
orld
)
For
a y
ear,
20%
dec
line
in a
ir tr
avel
and
in s
ervi
ces
sect
or.
Glo
bal e
cono
mic
link
ages
of t
rade
flow
adj
ustm
ent a
nd c
apit
al fl
ow r
eal-
loca
tion
.
Sim
ulat
ions
.
Bre
akdo
wn
ofth
e ec
onom
icim
pact
s (m
orta
l-it
y, m
orbi
dity
and
dem
and
chan
ges)
.
Glo
bal
1918
-like
pan
dem
ic s
cena
rio:
dev
elop
ing
coun
trie
s w
ould
be
hit t
heha
rdes
t (5.
3% fa
ll in
GD
P).
Hig
h-in
com
e co
untr
ies
4.7%
fall
in G
DP
.G
reat
glo
bal r
eces
sion
(4.
8% fa
ll in
GD
P).
Hum
an-t
o-hu
man
pan
dem
ic: t
otal
impa
ct 3
.1%
with
the
high
est i
mpa
ctco
min
g fr
om s
hift
s in
dem
and
(1.9
% o
f GD
P).
Initi
al im
pact
pur
ely
from
add
ition
al d
eath
s.
Pub
lican
d pr
ivat
e ef
fort
s to
miti
gate
the
spre
ad o
f the
dis
-ea
se b
y im
posi
ng tr
avel
res
tric
tions
and
soc
ial d
ista
nc-
ing
have
a la
rge
impa
ct o
n re
al a
ctiv
ity.
As
the
dise
ase
spre
ads
to th
e re
st o
f the
wor
ld, g
loba
lec
onom
ic a
ctiv
ity d
eclin
es s
igni
fican
tly in
an
effo
rt to
mon
itor
the
outb
reak
.
Fan
et a
l.(2
016)
E
stim
ates
of t
he p
roba
bilit
ies
of p
an-
dem
ics
on a
n an
nual
bas
is.
Tw
o se
veri
ty s
cena
rios
: mod
erat
e sc
e-na
rio
has
age-
spec
ific
mor
talit
y di
stri
-bu
tions
like
195
7 an
d 19
68 a
nd s
ever
esc
enar
io h
as a
ge-s
peci
fic m
orta
lity
dis-
trib
utio
ns li
ke 1
918.
Exp
ecte
d se
veri
ty is
def
ined
in te
rms
of s
tand
ardi
sed
mor
talit
y un
its.
Incl
usiv
e co
st
of a
pan
dem
ic.
Mod
erate scen
ario
Ann
ual e
xces
s m
orta
l-ity
rat
e in
the
low
er-
mid
dle
inco
me
coun
-tr
ies
is e
xpec
ted
to b
eab
out 0
.06%
(18
,000
deat
hs)
For
the
wor
ld, t
heex
cess
exp
ecte
d nu
m-
ber
of d
eath
s is
37,0
00.
Exp
ecte
d an
nual
in-
com
e lo
sses
are
USD
16bi
llion
.
Severe scena
rio
Ann
ual e
xces
s m
orta
l-ity
rat
e in
the
low
er-
mid
dle
inco
me
coun
-tr
ies
is e
xpec
ted
to b
eab
out 1
.2%
(37
0,00
0de
aths
).
For
the
wor
ld, t
heex
cess
exp
ecte
d nu
m-
ber
of d
eath
s is
680,
000.
Exp
ecte
d an
nual
in-
com
e lo
sses
are
USD
64 b
illio
n.
Total im
pact
Ann
ual i
ncom
e lo
ss:
1.6%
fall
in lo
wer
-mid
-dl
e in
com
e co
untr
ies
and
0.62
% fa
ll fo
r th
ew
orld
eco
nom
y.
Exp
ecte
d an
nual
incl
u-si
ve c
ost o
f 0.7
% o
fgl
obal
inco
me.
As
the
seve
rity
of t
he p
ande
mic
incr
ease
s, th
e in
trin
sic
cost
of p
rem
atur
e de
ath
and
illne
ss r
ises
, and
this
is fa
rm
ore
obvi
ous
for
the
low
er-m
iddl
e in
com
e co
untr
ies.
Assum
ptions
Mod
el/Aim
Scop
eGDP im
pact
Com
men
ts
Table
3Su
mm
ary
of p
aper
s w
hich
mod
el p
ast
pan
dem
ics
(continued)
and instrument Spanish Flu mortality with theprevious year’s influenza mortality for theregion. They find strong effects on manufac-turing employment and output, bank assets,and car registrations, pointing to both supplyand demand channels, as well as financial fric-tions.10 The estimates imply that the pandemicreduced manufacturing output by 18% forregions with average exposure,11 while nationalbanks saw an increase in losses charged off rel-ative to assets in 1920-21, indicating anincrease in NPLs in 1919-20.
Velde (2020), on the other hand, uses a vari-ety of high-frequency data to argue that theshort-term economic effects of the pandemicwere quite modest. Industrial output fellsharply but rebounded after a few months, thefinancial system was robust, and business fail-ures were minimal. Business failures did risesubstantially, and several measures of eco-nomic activity (including retail trade and pay-ments) contracted severely in the second halfof 1920, with industrial production reaching atrough in May 1921. However, the 1920 wavewas much smaller than the 1918-19 wave, andoccurred in February. As Velde (2020) notes,the discrepancy between his results and thoseof Correia et al. (2020) and Brainerd andSiegler (2003), who show negative effects onlong-term outcomes, presents a challenge foreconomics research, highlighting the need tofind a state variable that propagates the shockof 1918 to 1923 and later.
Correia et al. (2020) also consider the eco-nomic effects of NPIs. They build on the epi-demiology literature establishing that NPIsdecrease influenza mortality, and use variationin the timing and intensity of NPIs across citiesto study their economic effects. They find thatcities that intervened earlier on and moreaggressively experienced a relative increase inmanufacturing employment, manufacturingoutput, and bank assets in 1919, after the endof the pandemic. The effects are economicallysizeable. Reacting 10 days earlier to the out-break of the pandemic in a given city increasedmanufacturing employment by around 5% in
the post-pandemic period. Likewise, imple-menting NPIs for an additional 50 daysincreased manufacturing employment by6.5% after the pandemic. In 1919 and 1920, anincrease in banking assets was observed incities with early and longer interventions after1918, which helped to mitigate the exacerba-tion of the crisis that resulted from bankdeleveraging due to higher defaults. Alto-gether, their findings suggest that pandemicscan have substantial economic costs, and NPIscan have economic merits, beyond loweringmortality.
They conjecture that the results may be drivenby the fact that the pandemic itself can haveimportant economic effects, as people cutback on consumption and labour supply, andthat NPIs can reduce the length of disruptionby solving coordination problems. On theother hand, a particularly strong channel forthis pandemic was probably the fact that it tar-geted prime-age adults, and so NPIs hadstrong effects in preserving the local labourforce. An important channel of transmissionof both demand and supply shocks could havebeen the banking sector. The temporarynature of the pandemic should in principlelead to increased demand for liquidity (Holm-ström and Tirole 1998), and healthy bankscould then smooth the shock and mitigate thedecline in demand and production. Wide-spread defaults, however, may stress the bank-ing system, impairing its assets, and poten-tially amplifying the pandemic to a financialcrisis. This has important implications forCOVID-19; bridge loans to levered actors toprevent unnecessary defaults and destructionof productive capacity are key, and it is no sur-prise that they are an essential element of res-cue packages. Nevertheless, it should be noted
51Economic BulletinJuly 202016
10 Local manufacturing should be somewhat insensitive to changes inlocal demand, so lower relative manufacturing employment wouldbe indicative of a supply shock. The opposite holds for carregistrations, while for bank assets both types of effects arepossible; supply shocks may lead to defaults, while lower demandmay shrink lending. Credit rationing could be an importantamplifier of the shock.
11 A concern is that data on manufacturing outcomes are onlyavailable for 1914 and 1919, which makes it hard to control for theeffect of WWI. However, data on car registrations and bank assetsare annual, thereby sharpening identification substantially.
that Velde (2020) finds little short-termeffects of mortality on bank outcomes at thecity level, making the connection to long-runoutcomes puzzling.
On the other hand, Barro (2020) finds no evi-dence of a relationship between NPIs and mor-tality. Though the curve was flattened, in thatthe ratio of peak to average deaths did fall, thetotal effect was unrelated to NPIs. He arguesthat this is because the measures were notimplemented long enough to have substantialeffects, as they had an average duration ofaround one month. Yet it is possible that sometypes of NPIs may be more effective than oth-ers, as he finds significant negative effects forrestrictions on public gatherings.
Furthermore, Lilley et al. (2020) collect alarger sample of data and argue that the resultsof Correia et al. (2020) regarding the effects ofNPIs are driven by pre-existing trends in pop-ulation, and hence manufacturing employmentand output. In fact, they find that NPIs arestrongly related at the city level with popula-tion growth ten years before the pandemic,causing a spurious relationship between NPIsand employment growth. Once this is takeninto account, results are uninformative aboutthe true effects of NPIs.12
Barro et al. (2020) study the macroeconomicimpact of the 1918 influenza pandemic at thecountry level for 42 countries, separating theeffect of WWI by controlling for the deathsof soldiers in combat. The analysis yields flu-generated declines for GDP and consumptionin the typical range of 6%-8%, respectively.13
The results cannot rule out effects of the flupandemic on the level of real per capita GDPthat are fully permanent, or fully temporary,or somewhere in between. The authors alsoprovide some evidence that higher flu deathrates decreased realised real returns onstocks and, especially, on short-term gov-ernment bills. There is no prediction that theshort-term negative effect will be reversed.Finally, the results on inflation confirm thatthe 1918 influenza pandemic and, especially
WWI, increased inflation rates at least tem-porarily.
With regard to lessons for the COVID-19episode, it should be stressed that the indus-trial structure is substantially different nowthan a century ago. Notably, services nowaccount for a much larger share of the econ-omy compared with the late 1910s, whereasthe opposite holds for manufacturing or agri-culture. A large portion of services aredemanded at a specific point in time, with asmaller role for pent-up demand than is thecase with durable or even non-durable goods.For instance, restaurant meals foregone dueto closures will not be recovered once lock-downs are lifted (even abstracting from lowerdemand due to continued fear of infection orincome uncertainty), unlike purchases of dish-washers or furniture. As such, even if thedownturn resulting from the Spanish Flu wasshort-lived, this does not necessarily implythat the COVID-19 effects will follow a sim-ilar path.
Other interesting papers that study the effectof the Spanish Flu include Almond (2006) andLe Moglie et al. (2020). Almond (2006)explored a longer-term effect of the SpanishFlu: whether in utero exposure to the influenzahad negative economic consequences for indi-viduals later in their lives. The hypothesis isthat individuals’ health endowments are posi-tively related to human capital and productiv-ity and thus also to wages and income (the fetalorigins hypothesis). Using 1960-80 decennial
51Economic Bulletin
July 2020 17
12 In a later response to the findings of Lilley et al. (2020), Correiaet al. argue that the population values used by the authors areproblematic, as they are not census-based and reflect extrapolationsfrom values between 1900 and 1910, leading to measurement error.Furthermore, they argue that the spurious relationship betweenNPIs and employment growth documented in Lilley et al. (2020)is not present with employment growth five years before thepandemic. Accounting for population growth gives results close tothe original ones. See http://scorreia.com/research/pandemics-llr-response.pdf.
13 The flu death rate for 1918-20 has an overall correlation of -0.25with a country’s real per capita GDP in 1910. This may largelyreflect the impact of better health services and betterorganisation on the probability of death from the disease (reflectingpartly risk of infection and partly the mortality rate given infection).An offsetting force, however, is that more advanced economies arelikely to have greater mobility and interactions, which foster thespread of contagious diseases.
census data, Almond (2006) found that cohortsin utero during the 1918 pandemic hadreduced educational attainment, higher ratesof physical disability, and lower incomes. “Chil-dren of infected mothers were up to 15% lesslikely to graduate from high school. Wages ofmen were 5-9% lower because of infection”(Almond 2006, p. 673).
In a very recent paper, Le Moglie et al. (2020)study the social effects of the Spanish Flu. Theextremely high rates of infection and fatality(over 2%, relative to less than 0.1% for the typ-ical flu), together with public health guidelinesto limit physical contact, similar to the meas-ures taken against COVID-19, had a substan-tial effect on social interactions. Using a rep-resentative survey of the descendants of immi-grants to the US, they show that descendantsof immigrants coming from countries withhigher influenza mortality had significantlylower levels of social trust, with an additionaldeath per thousand population being associ-ated with a 1.4 percentage point decrease intrust, relative to descendants of immigrantswho had migrated from the same countriesbefore the Spanish Flu. A possible mechanismis the comparison between countries that wereneutral and belligerent in WWI; in the former,lack of censorship meant that societies inter-
nalised the threat of the disease and the needfor social distancing.
Some studies have also looked at morelocalised pandemics, such as the Severe AcuteRespiratory Syndrome (SARS) outbreak inAsia, which caused economic disruption eventhough a global health crisis was averted. Stud-ies of the macroeconomic effects of the SARSepidemic in 2003 found significant effects oneconomies through large reductions in con-sumption, an increase in business operatingcosts, and a re-evaluation of country risksreflected in increased risk premia (for a review,see McKibbin and Fernando 2020). Shocks toother economies were transmitted according tothe degree of the countries’ exposure to thedisease and to the affected economies. TheAsian Development Bank estimated that theeconomic impact of SARS amounted toaround USD 18 billion, or 0.6% of GDP, inEast Asia (Fan 2003), mainly through itseffects on consumption.14
51Economic BulletinJuly 202018
14 Other studies of SARS include Chou et al. (2003) for Taiwan, andSiu and Wong (2004) for Hong Kong. These studies focus mostlyon assessing the damages induced by SARS in affected industriessuch as tourism and the retail services sector. Siu and Wong (2004)reported that retail sales figures dropped by 15.2% between late2002 and April 2003. They also reported substantial declines of10.4% in passenger travel to Hong Kong over a similar timeperiod.
Impulse responses to major pandemics
A. Response of the European real natural rate of interest following pandemics
Source: Jordà et al. (2020).
B. Response of real wages in Europe following pandemics
Note: Response of the real natural rate of interest and of real wages to a pandemic in Europe, one to 40 years into the future. Shaded areas are 1 and 2 s.e. bands around response estimates.
Finally, Jordà et al. (2020) study the effect ofpandemics across all major events since theBlack Death, looking at outcomes up to 40years out. They study rates of return onassets15 using a dataset that covers France,Germany, Italy, the Netherlands, Spain, andthe UK, focusing on 15 major pandemicepisodes, where more than 100,000 peopledied. The results show that following a pan-demic, the natural rate of interest declines fordecades thereafter, reaching its nadir about 20years later, with the natural rate about 150basis points lower than what would have beenthe case if the pandemic had not taken place,and returns to trend around four decades later(see panel A of the chart). The heterogeneityof the responses is quite striking. For France,Italy, and Spain the effects of pandemics aremuch larger (3%-4%) relative to Germany,the Netherlands and the UK. This reflects,among other explanations, the timing of thepandemics across countries, the relative expo-sure, the relative size of the working popula-tion, and the relative degrees of industriali-sation. The authors also look at some morelimited evidence on real wages and find thatthe response of real wages is almost the mir-ror image of the response of the natural rateof interest, with its effects being felt overdecades, a result consistent with the baselineneoclassical model (see panel B of the chart).Real wages gradually increase until aboutthree decades after the pandemic, where thecumulative deviation in the real wage peaks atabout 5%. All results are robust to controllingfor wars, possible major trend breaks and afteromitting the Black Death and the Spanish Flu.The results indicate that pandemics are fol-lowed by decades of depressed investmentopportunities, possibly due to excess capitalper unit of labour. The authors also speculatethat there may also be a higher propensity tosave (due to higher risk aversion or a rebuild-ing of depleted wealth), but they do not reallyprovide any evidence for this. Instead, theyshow that wars are associated with an increasein natural rates, suggesting that sharpchanges in capital per unit of labour maytrump demand effects.
2.2 MODELLING THE ECONOMIC IMPACT OF PASTPANDEMICS
Meltzer et al. (1999) examine the likely eco-nomic effects of an influenza pandemic in theUS and evaluate several vaccine-based inter-ventions. They use a Monte Carlo mathemat-ical simulation model with predefined proba-bility distributions for a set of input variablesby age and risk group. At a gross attack rate(i.e. the number of people that become clini-cally ill out of the total population) of 15%-35%, the estimated total economic impact (i.e.disease-associated medical costs, indirect costsfrom losses of time and income by careers, andmorbidity/mortality costs measured asexpected foregone lifetime earnings) for theUS economy ranged from USD 71.3 billion toUSD 166.5 billion (excluding disruptions tocommerce and society), and the estimateddeaths ranged from approximately 89,000 toapproximately 207,000 people. The costs asso-ciated with mortality accounted for approxi-mately 83% of all costs. Thus, the results indi-cate that in the event of a pandemic any inter-vention should aim at lowering the death rates.The authors also argue that vaccinating prior-ities should be set depending on the policyobjectives (preventing deaths, maximising eco-nomic returns, or other), leading to differentvaccine-based intervention plans. However, theresults for the economic returns of vaccinationschemes presented in the paper are sensitive tothe assumed gross attack rate and cost of vac-cination. Furthermore, the authors use an eco-nomic value of life of USD 1 million, whileconventional current estimates put it at aroundUSD 10 million for the US, implying that theseestimates may be an order of magnitude higherin current US dollars. Finally, the authorsacknowledge that (a) other multiplier effectsresulting from disruptions in commerce andsociety need also to be valued and incorpo-rated in the model, and (b) the range of the
51Economic Bulletin
July 2020 19
15 Aggregate real interest rates are constructed by weighting realinterest rates on long-term debt by GDP shares (Maddison 2010).The underlying assets are debt contracts, “which are not contractedshort-term, which are not paid in-kind, which are not clearly of aninvoluntary nature, which are not intra-governmental, and whichare made to executive political bodies”.
gross attack rate of an actual pandemic is quitewide, thereby leading to considerable uncer-tainty about a pandemic’s potential economicimpact.
Bloom et al. (2005) use the Oxford economicforecasting global model to estimate the poten-tial short-run economic impact for Asia of apandemic resulting from a mutation in theavian flu that spreads from human to human.They look at two scenarios, both assuming amild pandemic spread around a year with a20% attack rate and a 0.5% case-fatality rate.Under the first scenario, there is a consump-tion shock of 3% and a reduction in the tradeof services, which last for two quarters and onlyaffect Asian countries. Supply shock isassumed as a two-week absenteeism fromwork. The economic impact for Asia is esti-mated at around USD 99.2 billion from thedemand side and USD 14.2 billion from thesupply side. Under the second scenario, wherethe shocks last for four quarters, the impact isgreater since there is also a reduction in con-sumption in the rest of the world, leading to aglobal GDP shrinkage of 0.6% and a contrac-tion of 14% in global trade of goods and serv-ices. This implies that open economies andexporters of services are more vulnerable tointernational shocks. For Asia, GDP growth ispredicted to remain low even five years afterthe pandemic. Taking all these together, theresults show that timely policy responses canhelp prevent and mitigate the economic impactof a pandemic, also necessitating cooperationand coordination among countries.
Lee and McKibbin (2004) use a global model(called “G-Cubed Asian Pacific” model), con-sisting of 20 countries and 6 sectors, to esti-mate the impact of SARS arguing that (a) direct medical costs and demographiceffects because of the epidemic might be low(or at least ambiguous for the supply-sideeffect), and (b) there are important linkagesbetween and within economies through bothinternational trade and capital flows. Hence, aglobal model that takes into account integratedeconomies, rational expectations and forward-
looking intertemporal behaviour (although thisis acknowledged to be an unrealistic assump-tion for the real world when a new diseaseappears) is more adequate to get a full pictureof the channels of transmission and of the eco-nomic costs from a global disease. Under theassumption of a temporary shock (captured asan increase in country risk premium, a drop indemand in the services sector and an increasein costs in the services sector) of six months tothe affected economies, and by calculating acountry-specific index of “global exposure toSARS” (depending, among other things, ongeographical distance to China and governanceresponse) to scale the shock to other countries,the authors find that despite a relatively smallnumber of cases and deaths, the global costs ofSARS were significant and not only limited tothe countries directly affected. They estimateGDP losses of between 2.63% and 0.5% forChina, Hong Kong, Singapore and Taiwan.Analysis of the data revealed that the retailsector and travel industries suffered the largestdeclines. The model also predicts how theexpectations about future developmentsrelated to the disease might affect the poten-tial economic costs of SARS. The more per-sistent SARS is assumed to be (lasting up toten years), the larger the negative economicimpact in affected economies, but the smallerthe impact in the rest of the world, whichreflects the direction of capital flows to theleast affected countries. The calculations sug-gest that the cost of SARS in 2003 for theworld economy was close to USD 40 billion inthe event of a temporary SARS effect andaround USD 54 billion in the event of a per-sistent SARS shock (not including actual costsin later years if in fact SARS does persist). Therecession is found to last for a number of yearsafterward. Under a persistent shock, for theaffected economies, the primary impactcomes from the rise of the country risk pre-mium, leading to a sharp decrease in invest-ment. Net capital outflows from China andHong Kong are estimated to 0.8% and 1.4%,respectively, with a positive effect on theirtrade balances due the exchange rate depreci-ation. The results overall point towards the
51Economic BulletinJuly 202020
argument that in a complex interrelated world,even if toll rates are low and demographiceffects are insignificant compared with othermajor epidemics, a disease outbreak mighthave a huge economic impact for the globaleconomy and not only for the affectedeconomies.
McKibbin and Sidorenko (2006) adapt the Leeand McKibbin (2003) G-Cubed Asian Pacificmodel (also extended here to include the UK)to examine the global economic consequencesof a range of pandemic influenza severities: amild pandemic similar to the 1968-69 HongKong flu, a moderate pandemic similar to the1957 Asian flu, a severe pandemic similar tothe 1918-19 Spanish Flu and an “ultra” pan-demic scenario (toward the upper end of therange of estimates for severity in 1918), whichis not based on any known previous pandemicbut has the characteristics of the Spanish Fluin addition to higher mortality rates for olderpeople. In the model, the various epidemicshocks (estimated using some indices captur-ing the possibility for each country that eachseverity might occur and a sickness index thatcaptures morbidity rate), given an attack rateof 30%, affect Asian economies through alarge reduction in consumption (modelled asendogenous shifts), an increase in the cost ofdoing business (scaled for the services sectorexposure of the economy across countries andfor the mortality shocks across scenarios) andan increase in country risk premia (calculatedas a composite indicator of the responsivenessof the health services sector to the pandemic,the quality of governance and the exposure ofthe country to foreign capital). At the sametime, shocks are transmitted to othereconomies depending on the global exposureof the country to the disease. Shocks are shownto be temporary, lasting only for 2006 for mostcountries and fading out until 2008 when realactivity recovers. The authors estimate that theultra scenario would lead to over 142.2 milliondeaths, and to income losses of over 12% ofGDP (USD 4.4 trillion) worldwide and over50% in some developing countries such asHong Kong in 2006. Even under a mild sce-
nario it is estimated that a pandemic wouldcost 1.4 million lives worldwide and globalGDP would fall by approximately 0.8%, withmortality and morbidity shocks being the maindrivers, assuming that monetary policy caneffectively contaminate demand changes. Asthe severity of the pandemic increases, theimportance of cost rises as well, resulting in alarger shrinkage of global GDP, but with thenegative effect being stronger for Asian anddeveloping countries, where contraction ofdemand is larger, mortality rates are higherand capital outflows are more substantial.Under the severe scenario, contraction in somecountries of Asia reaches 26% relative to thebaseline scenario of no pandemic, partlyreflecting large reallocation of capital to more“safe” economies. The composition of theresults shows a great difference among coun-tries, with developing countries being the mostadversely affected. Although prices rise in theshort run, the result depends on whether thedemand-side or the supply-side effect isstronger. Monetary tightness (as a response todeclining output, inflation changes or exchangerate changes) results in a more severe eco-nomic impact and may have a great relevancefor bond markets, together with fiscalresponse. The paper concludes that althoughthere is high uncertainty about how aninfluenza pandemic evolves with little histor-ical guidance on how people tend to reactunder a severe situation and how such a crisisshould be managed, policy makers shouldinvest a lot in averting an outbreak because ofits significant potential consequences for theglobal economy (see Table 4).
Previous studies by the World Bank (see Jonas2013) looked at pandemic risk and estimatedthat a 1918-like pandemic could cost USD 3trillion globally. Burns et al. (2006) replicatethe results of McKibbin and Sidorenko (2006)and find that, under a 1918-like pandemic sce-nario (severe scenario), developing countrieswould be hit the hardest (5.3% fall in GDP)relative to the high-income countries (4.7% fallin GDP) and there would be a great globalrecession (4.8% fall in GDP). The authors also
51Economic Bulletin
July 2020 21
present results under a different scenario of ahuman-to-human pandemic with a similar mor-tality rate to the Spanish Flu (1.08% of peopledie across the world) and provide a breakdownof the economic impacts from mortality, mor-bidity and demand changes (where it isassumed a 20% decline in air travel and inservices). The total economic impact is a 3.1%fall in world GDP, with the strongest impactcoming from shifts in demand (1.9% of GDP),i.e. public and private efforts to mitigate thespread of the disease by imposing travel restric-tions and social distancing have a large impacton real activity. The main policy implication ofthe simulation study is that there is a big imme-diate impact on the affected countries, but asthe disease spreads to the rest of the world,global economic activity declines significantlyin an effort to contain the outbreak (see Table5).
In a different strand, Fan et al. (2016) assessthe inclusive cost of a pandemic that adds tothe income loss the intrinsic value of prema-
ture mortality and morbidity. They use esti-mated probabilities of pandemics on an annualbasis under two severity scenarios to provideestimates of mortality rates and of the associ-ated cost. Expected severity is defined in termsof standardised mortality units16 and it isassumed that the moderate scenario has age-specific mortality distributions like 1957 and1968 and the severe scenario has age-specificmortality distributions like 1918. The resultsshow that the annual excess mortality rate dueto a pandemic in the lower-middle incomecountries is expected to be about 0.06%(18,000 deaths) under the moderate scenarioand about 1.2% (370,000 deaths) under thesevere scenario. For the world, the excessexpected number of deaths is 37,000 and680,000 people, respectively. Translatingthese figures into an annual income loss givesa 1.6% fall in lower-middle income countriesand a 0.62% fall for the world economy
51Economic BulletinJuly 202022
Average Annual Average Annual
Australia 5.16 -7.70 0.89 -1.84
Canada 8.69 -6.70 3.54 -11.14
United Kingdom 2.88 1.00 3.02 -13.89
Japan 2.52 -3.00 5.56 7.94
United States 5.48 -7.70 6.09 -5.22
SPANISH FLU, 1918-1919
GDP growth (%) 1908-13 1914 1914-18 1919
Average Annual
Australia 4.38 4.80
Canada 5.28 1.79
United Kingdom 3.06 -0.21
Japan 8.43 5.83
United States 2.62 -0.49
ASIAN FLU, 1957-1958
GDP growth (%) 1953-57 1958
Table 4 Summary of estimated effects from two recent pandemics
Source: McKibbin and Sidorenko (2006).
16 The standardised mortality unit (SMU) represents a 10-4 mortalityrisk and is used to represent small numbers as integers.
(annual mortality cost). For the world, theexpected annual income losses amount to USD16 billion under the moderate pandemic andUSD 64 billion under the severe pandemic sce-nario. This means that as the severity of thepandemic increases, the intrinsic cost of pre-mature death and illness rises as well, and thisis far more obvious for the lower-middleincome countries. Adding these two coststogether gives an expected annual inclusivecost of 0.7% of global income.
3 UNDERSTANDING THE ECONOMIC IMPACT OFCOVID-19
3.1 MODELLING COVID-19: CONVENTIONALECONOMIC MODELS
Motivated by the recent outbreak of COVID-19, a number of papers use full-scale macromodels to gauge its potential economic impactand explore the relative merits of alternativeeconomic policy responses. According to theliterature, the economic impact of the currentpandemic can be thought of as entailing botha supply shock, stemming from the self-imposed social distancing and the state-imposed lockdown, and a demand shock, i.e. anegative shock to consumption resulting fromreduced opportunities to consume. Moreover,the pandemic increases risk and uncertainty
both in the real economy and in the financialsector. However, modelling a pandemic toexamine its impact on the economy is chal-lenging, due to the large degree of uncertaintywith respect to the nature, the persistence andthe size of the shocks arising from the pan-demic. Indeed, different papers adopt very dif-ferent approaches.
Fornaro and Wolf (2020) choose to model theimpact of the pandemic as a drop in labourproductivity growth. They opt for an unex-pected, very highly persistent shock, as theywant to explore the pessimistic scenario thatthe COVID-19 outbreak leads to a long-last-ing supply disruption. However, they point outthat the persistence of the shock (supply dis-ruption) is crucial to their analysis, as itinduces agents to revise their expectations offuture income downwards.17 They find thatsuch a shock leads to a decline in demand and,given the downward revision of agents’ futureincome expectations, to a decline in output andemployment. If we assume that the decline indemand also leads to a decline in investment,
51Economic Bulletin
July 2020 23
17 In earlier theoretical work, Torój (2013) had proposed that, withina standard DSGE model, an epidemic can be modelled as areduction in labour utilisation under unchanged labour cost. Sucha shock has a smaller impact than a negative labour supply shock,as the latter directly affects the nominal wage. It is argued that sucha representation is adequate when computing the costs of short-lived diseases like epidemics of influenza, if one assumes away anypossible long-run consequences stemming from higher morbidityor mortality.
World -0.7 -2.0 -4.8
High-income countries -0.7 -2.0 -4.7
Developing countries -0.6 -2.1 -5.3
East Asia -0.8 -3.5 -8.7
Europe and Central Asia -2.1 -4.8 -9.9
Middle East & North Africa -0.7 -2.8 -7.0
South Asia -0.6 -2.1 -4.9
Deaths (millions) 1.4 14.2 71.1
Mild Moderate Severe
Table 5 Possible economic impacts of a flu pandemic
(% change in GDP, first year)
Source: Burns et al. (2006).Note: The mild scenario is modelled on the Hong Kong flu of 1968-69; the moderate flu has the characteristics of the 1957 Asian flu; and thesevere scenario is benchmarked on the 1918-19 Spanish Flu.
the initial shock may generate a “supply-demand doom loop”, as a decline in investmentendogenously generates a further drop in pro-ductivity. Within this model, monetary stimu-lus can mitigate the impact of COVID-19 onemployment and output, and will have a mul-tiplicative positive effect by containing thedoom loop. If, however, monetary policy is lim-ited by the zero lower bound, then agents’expectations of lower income become self-ful-filling and may lead to a stagnation trap. In thiscase, fiscal policy should be used to supportpublic and private investment, thus helpinglabour productivity (and expectations) torebound. In this context, it is unclear whetherinflation will be pushed upwards or down-wards, but the central bank does face a trade-off between stabilising inflation and employ-ment.
Bayer et al. (2020) assume that a random frac-tion of workers (10% or 30%, depending on thescenario) is in quarantine, i.e. they have zeroproductivity, a state from which they recoverquickly (there is a 50% exit probability at theend of each period). The shock leads to areduction in output, while its randomnessintroduces income risk into the model, as it isex ante unknown exactly which households willbe quarantined and which will not. The setupis a HANK-DSGE model with numerous fric-tions and incomplete financial markets, due towhich idiosyncratic risk is uninsurable at thehousehold level. The authors find that thequarantine shock, which reduces the effectivelabour force, leads to a sharp decline in con-sumption and investment upon impact, andthus to a reduction in output. The increase inincome risk leads to a decline in aggregatedemand and an increase in households’ pre-cautionary cash holdings, as they try to self-insure. The decline in demand prompts adecline in investment and thus a further declinein output. In other words, the ensuing recessionis due to both depressed supply and depresseddemand. Output falls by a maximum of 3.5% inthe trough period Q3, in the scenario where10% of the labour force is in quarantine for onequarter, and by a corresponding 11% if 30% of
the labour force is in quarantine. Recovery isslow – it takes up to three years to fully recoverfrom the shock, as some of the job losses arepersistent. The authors explore the potentialimpact of the recently announced US fiscalpackage and find that the transfer componenthas a high multiplier (between 0.4 and 1.2),depending also on the responsiveness of mon-etary policy to inflation. Key for this high mul-tiplier are the transfers which are directly paidto the unemployed and/or the quarantinedhouseholds, as these have a high marginalpropensity to consume, and the transfers exante reduce their income risk.
Guerrieri et al. (2020) assume that a fractionof agents are “unable” to work in the firstperiod, because the epidemic has renderedtheir occupation unsafe. They explore the cir-cumstances under which a negative supplyshock, such as one induced by a pandemic, cangenerate a demand shortage and thus an out-put decline larger than the initial shock – a typeof shock which they term “Keynesian” supplyshock. The authors show that, in the modelspecification with two sectors and incompletemarkets, when the initial supply shock only hitsone sector, such “Keynesian” supply shockscan materialise. They also model a “businessexit multiplier”, which occurs when firms areput out of business due to the epidemic,prompting a cascade of “business exits”, astheir employees cut back on their consumption– a second type of “Keynesian” supply shock.Low substitutability across sectors and incom-plete markets with liquidity constrained con-sumers, all contribute towards the possibilityof such shocks. Fiscal policy is overall lesseffective in such a model. However, optimalpolicy, which comprises a containment of theepidemic, a loosening of monetary policy anda social insurance policy which compensatesworkers in the affected sectors, alleviates thepossibility of a “Keynesian” supply shock andmakes it easier for public health objectives tobe pursued.
Faria-e-Castro (2020) models the COVID-19pandemic as a demand-side shock, i.e. a large
51Economic BulletinJuly 202024
negative shock to the marginal utility of con-sumption that induces households to reducecurrent consumption. The shock is imposedon the sectors that produce contact-intensiveservices – according to the author, these com-prise hospitality and leisure, certain types ofretail (bricks and mortar) and air travel.Additionally, it is assumed that the output ofthese sectors is unlikely to be consumed bythe government so that, in the model, a risein government spending does not boostdemand for contact-intensive services. Theauthor calibrates the size of the shock so thatthe rate of unemployment rises to 20%, tocapture the worst-case scenario put forwardby the Treasury Secretary to Members of theUS Congress on 17 March 2020. He assumesthat the pandemic lasts exactly three quarters,i.e. that there is a shock of equal size in eachquarter and that the economy returns imme-diately to the initial state once the pandemicis over. Constrained agents (modelled as bor-rowers), who have a high marginal propensityto consume, will be affected and thus non-ser-vice consumption will also decline. Con-strained agents will default on their loansmore often, leading banks to charge higherinterest rates on loans, thus further depress-ing overall consumption, demand for labourand inflation. Expansionary monetary policyhelps, but if constrained by the zero lowerbound, a deep recession can ensue. By design,the economy rebounds immediately after theshock. The author considers one-quarter fis-cal interventions of five alternative types(increase in government consumption, cut inlabour income tax, increase in unemploymentinsurance, unconditional transfers to allagents and a per-wage subsidy to services sec-tor firms) designed to have a comparableimpact on the fiscal balance. The latter inter-vention is the only one which protectsemployment in the services sector. The meas-ure that yields the highest GDP multiplier isgovernment consumption. However, theauthor acknowledges that there may be strongcomplementarities between policies, whichare not further examined.
McKibbin and Fernando (2020) use a globalmodel (a hybrid between a DSGE model anda computable general equilibrium model)which explicitly models 20 countries and 4sub-regions for the rest of the world and 6 sec-tors, with cross-country trade linkages. In linewith similar work of theirs for SARS, theyconsider seven epidemiological scenarios:three scenarios where the epidemic only hitsChina (with varying mortality and morbidityrates) and affects the rest of the world only viachanges in trade, capital flows and risk pre-mia; three scenarios of a pandemic affectingall countries; and one scenario of a mild annu-ally recurrent pandemic. These epidemiolog-ical scenarios are then mapped into the fol-lowing economic shocks:
• A labour supply shock which comprises amortality rate and a morbidity rate. Themortality rate is set on the basis of data onthe SARS epidemic (0.02%-0.9% dependingon the scenario). The morbidity rate reflects(i) the share of the population that will con-tract the virus and will have to stay at homefor 14 days (1%-30%) and (ii) the notionthat for every sick person a carer will alsotake sick leave, while 70% of the femalelabour force participation will have to stayat home for 14 days due to school closures.This is then adjusted from country to coun-try based on indices reflecting the extent oflinkages to China, urban population densityand health security inter alia (adjustmentsbenchmarked against China). In the pan-demic scenarios, China suffers a shock to itslabour supply ranging between -1.05% and-3.44% (annualised), which implies a shockto labour supply of about half the size forEuropean countries.
• A shock to the equity risk premium which,for the milder scenario, is calibrated on thebasis of the US equity markets’ observed ini-tial response to COVID-19. This is thenused as a benchmark and adjusted for dif-ferent scenarios and for different countriesbased on country-specific indices of country
51Economic Bulletin
July 2020 25
risk, financial risk, governance risk, healthpolicy, etc.
• A shock to the cost of production (beyondlabour inputs) in each sector and country,meant to reflect the fact that trade as well asland, air and sea transport have beenaffected (though this is purely a cost effect,rather than a quantity effect). The authorscalculate the cost input of these sectors tothe six aggregate sectors in the model. Theybenchmark the shock for the mild scenarioto the percentage increase in the cost of pro-duction observed in the Chinese manufac-turing sectors during SARS. This is thenscaled across sectors and countries depend-ing on how exposed they are to China and toland, air and sea transport.
• A shock to consumption (over and above thedecline in consumption which stems fromdepressed income), due to changes in pref-erences, benchmarked against the reductionin consumption expenditure observed inChina over the SARS epidemic.
The authors calculate that the impact on GDPfor the euro area ranges from -2% to -8.5%approximately. The pandemic causes a sharpdrop in consumption and investment which,combined with the risk shocks, leads to a damp-ening of economic expectations, a sharp dropin equity prices and a move towards safe-havenbonds and cash, despite an endogenous easingof monetary policies. Capital flows out ofseverely affected economies like China andother emerging market economies and devel-oping countries and into safer advancedeconomies which experience a currency appre-ciation. This generates a corresponding adjust-ment of current accounts: countries whichexperience a capital flight see higher exportsand lower imports, while the trade balance ofadvanced economies deteriorates. The recov-ery is V-shaped in all countries and scenarios,except for the recurrent outbreak scenario.
A similar but much simpler exercise is under-taken by Luo and Tsang (2020), who first cal-
culate the impact of the loss of labour input inHubei province on its own production and thatof any other province in the country; then,using input-output tables and the industrycomposition of each province, they back outthe loss in aggregate output due to this loss oflabour input. Subsequently, they attempt toestimate the loss in global output based onglobal trade linkages.18 Luo and Tsang (2020)calculate that China will suffer an output lossof about 4% per month of lockdown, whileglobal output will correspondingly drop by 1%due to the economic contraction in China.About 40% of the impact is indirect (ratherthan a direct result of lower labour supply inthe affected region) coming from spilloversthrough the supply chain inside and outsideChina.
3.2 MODELLING COVID-19: ECONOMIC MODELSWITH AN EPIDEMIOLOGICAL BLOCk
An entirely new and innovative strand of eco-nomic literature has emerged, as academicsstrive to understand the impact of theCOVID-19 pandemic on economic variables.In a seminal paper, Eichenbaum et al. (2020)were the first to explicitly model the interac-tion between economic decisions and the rateof infection by embedding a standard epi-demiological model within a DSGE frame-work, i.e. by explicitly modelling the proba-bilities that economic agents transitionbetween the states: susceptible, infected, andeither recovered or “removed” (the SIR-macro model). The agents’ optimising behav-iour in each of these states is also explicitlymodelled. Both consumption and labour sup-ply increase the agents’ individual probabilityof infection (as well as the overall infectionrate) and thus economic agents optimallyreduce their labour supply and consumption– that is, without the introduction of a shockas in previous models. The economic and epi-
51Economic BulletinJuly 202026
18 In the same spirit, Fernandez (2020) also attempts to calculate theeconomic impact of COVID-19. For example, under theassumption of a 3-month lockdown, Greece is projected to suffera contraction of 6.5% in GDP. However, no details of the modelor the calculations are given.
demiological implications of alternative sce-narios are then examined.19
The agents’ micro-founded behaviour reducesthe economy-wide rate of infection and deathand generates a sharp recession, which peaks 32weeks from the outbreak at a maximum devi-ation from the steady state of 8% (the baselinescenario). If there is concern that the health-care system may become overwhelmed (mod-elled by making the mortality rate an increas-ing function of the number of peopleinfected), people cut back on work and con-sumption even more aggressively, as they self-impose stricter social distancing. In this case,the timing of the trough is earlier, but it is muchdeeper (reaching -22% on the trough week)and somewhat more protracted than in thebaseline scenario. From an epidemiologicalpolicy perspective, it is optimal to graduallyimpose compulsory social distancing measuresand to tighten them whenever infection ratesincrease and reduce them whenever infectionrates decline, until a critical share of the pop-ulation achieves immunity (optimal contain-ment scenario). Such a policy results in a reces-sion which peaks a few weeks later than in thebaseline scenario but is extremely protracted,reaching a maximum deviation of approxi-mately -20% from the steady state on thetrough week. When the model instead incor-porates a positive probability of a vaccine beingdiscovered, it is optimal from an epidemiolog-ical perspective to immediately introducesevere social distancing measures and to keepthem in place until the vaccine arrives, result-ing in an immediate (i.e. not bell-shaped) reces-sion, the trough of which is more protracted butless deep (maximum deviation from the steadystate at 14%). It should be noted that, in termsof the long-run post-epidemic equilibrium, theimpact of social distancing is positive: whileboth consumption and labour supply are lowerin the new post-epidemic equilibrium, theirdecline is smaller because a smaller cumulativenumber of deaths has been achieved.
A number of papers build on the seminal workof Eichenbaum et al. (2020). Jones et al. (2020)
extend the model by adding (i) multiple con-sumer goods with different contagion risk and(ii) the possibility of working from home whichrequires learning-by-doing. They find two addi-tional effects. First, the fatalism effect,whereby if agents think they will inevitably con-tract the virus, they may feel it is best to con-tract it early on. Second, the front-runnereffect, which also yields the same result: eco-nomic agents who worry that the healthcaresystem may become overwhelmed at a laterdate may opt to get infected today to ensurebetter healthcare. Both effects imply that opti-mal government interventions should be morefront-loaded.
Krueger et al. (2020) introduce different infec-tion probabilities across sectors but assumethat it is easy to substitute contact-intensiveservices (e.g. having a pizza at a pizzeria) withequivalent non-contact intensive ones (e.g.takeaway pizza). Similarly, they assume thatworkers are able to quickly relocate to sectorsnow in demand, e.g. waiters will do deliveries.Under this specification, the “Swedishapproach” to the epidemic, which prescribes nogovernment intervention, suffices to mitigateup to 80% of the human and economic cost. Inother words, in their model, endogenous shiftsin private consumption across sectors act as amitigation mechanism during the epidemic.However, the authors acknowledge that theirfindings crucially hinge on both substitutabil-ity and labour market flexibility.
Glover et al. (2020) model a number of dis-tinctions between agents: (i) young workers
51Economic Bulletin
July 2020 27
19 A number of recent papers explore variations and extensions of thestandard SIR epidemiological model which underlies the economicmodel of Eichenbaum et al. (2020); see for example Atkeson(2020), Berger et al. (2020), Casares and Khan (2020), Fergusonet al. (2020) and Stock (2020). However, they do not embed theseepidemiological models within a macroeconomic model and thusdo not explore the economic implications of the pandemic. Rather,their focus is on healthcare cost dynamics, forecasting the durationof the pandemic, exploring the optimal length of the lockdown andconsidering possible alternative approaches to testing – e.g. broadertesting of asymptomatic patients coupled with a more limitedlockdown and thus a smaller economic fallout. An exception ofsorts is Greenstone and Nigam (2020) who employ the Fergusonet al. (2020) simulation model for the US to estimate the numberof lives saved as a result of social distancing and then attempt to“monetize” them. This provides a quantification of the socialdistancing measures’ economic benefits.
versus old pensioners; (ii) healthy versus sickworkers; and (iii) basic sector (which cannot bequarantined) versus luxury sector (which can).They highlight the fact that the pandemicaffects these cohorts differently; to some theepidemic poses a major health risk, while notto others. Similarly, mitigating measures suchas the quarantine have a heterogeneous impacton agents: some will lose their income due tothe quarantine, while others won’t. In sum,both the pandemic itself and its mitigationhave distributional effects, as gains and lossesare unequally distributed. Mitigation measuresgenerate a need for redistributive policies, i.e.a need to tax agents who still work in order tocompensate those who cannot work for theincome lost. The authors show that optimalredistribution and mitigation policies interact,e.g. governments which find redistributionmeasures too costly may decide that a lowermitigation effort is optimal. Ultimately, theoptimal policy will reflect a compromisebetween the policy paths preferred by differ-ent subgroups of the population.
Finally, Velasco and Chang (2020) focus onthe dilemma of the healthy: to forego today’sincome by adhering to the quarantine in orderto be healthy tomorrow and enjoy tomorrow’sincome, or to work today and risk foregoingtomorrow’s income in case they get infected.They find that the initial income level mattersfor this choice: poorer workers are unlikely towillingly forego today’s salary. Thus, quar-antines are more difficult to enforce in poorereconomies than in advanced ones. However,economic policy can affect this choice. If thepolicy maker compensates agents for theirincome loss, they will be willing to complywith social distancing measures. Alterna-tively, the policy maker may commit to imple-menting expansionary policy in the nextperiod, so that agents find it optimal toadhere to social distancing measures today inorder to enjoy the next period’s higherincome. In all cases, policy credibility is cru-cial. The authors conclude that economic pol-icy can change the contagion dynamics via itsimpact on incentives.
It should be noted that most of the aforemen-tioned papers were written with the aim ofunderstanding the immediate implications ofCOVID-19 for the economy, i.e. with a short-to medium-term perspective in mind. Indeed,they barely touch upon the potential long-runeconomic implications of the COVID-19 pan-demic. Furthermore, in most of the papersmentioned above that use DSGE models, thepandemic shocks are temporary. Thus, theeconomy eventually returns to its pre-shocklong-run equilibrium. A notable exception arepapers in the spirit of Eichenbaum et al. (2020)where, in the long run, there is a permanentdecline in the labour force, equal to the cumu-lative number of deaths, leading to a perma-nent reduction in GDP. All in all, this litera-ture suggests that the duration of the downturndepends crucially on the duration of the lock-down, the degree of persistence of the result-ing economic shock and, where applicable, themortality rate. However, the authors acknowl-edge that they abstract from a number ofpotentially important determinants of an epi-demic’s economic impact, such as hysteresiseffects from unemployment, protracted bank-ruptcy costs and the destruction of supplychains inter alia, all of which could affect thelong-run performance of the economy andhave positive and normative implications.These and other economic aspects of COVID-19 are discussed in a non-technical manner intwo recent VoxEU/CEPR e-books (Baldwinand di Mauro 2020a and 2020b).20
3.3 THE IMPACT OF THE COVID-19 PANDEMIC ONECONOMIC ExPECTATIONS
The COVID-19 pandemic may also be affect-ing the economy through its impact on eco-nomic expectations, an avenue not directlyexplored in the aforementioned literature. Fet-
51Economic BulletinJuly 202028
20 The COVID-19 outbreak has also prompted new work on themeasurement and timing of economic activity. Leiva-Leon et al.(2020) propose an empirical framework for measuring the degreeof economic weakness of the global economy in real time and useit to gauge the impact of the COVID-19 pandemic. Laeven andValencia (2020) have updated their systemic banking crisesdatabase to facilitate comparisons between the current crisis andpast ones, improve understanding of how economically damagingthis crisis may be and inform policy making.
zer et al. (2020) use experimental data, surveydata and internet search data to measureCOVID-19 perceptions and explore how theyaffect economic expectations.21 They documenta rapid surge in economic anxiety since theCOVID-19 outbreak. They find that the exper-iment participants’ beliefs regarding mortalityrates and contagiousness causally affect theiranxiety regarding both the aggregate economyand their personal economic situation. How-ever, the participants’ aforementioned beliefsexhibit substantial heterogeneity and oftengrossly overestimate both contagiousness andmortality rates, thus potentially affecting theireconomic decisions disproportionately. Bartiket al. (2020) undertake a similar survey explo-ration of small businesses and find that firmstoo have widely varying beliefs about the likelyduration of COVID-19-related disruptions,while they also tend to anticipate problemswith accessing state aid, such as bureaucratichassles and difficulties establishing eligibility.These papers highlight the economic impor-tance of clearly and effectively conveying to thepublic the scientific facts on COVID-19, aswell as the need for timely policy measureswhich will both decrease economic hardshipand reduce perceived economic uncertainty, soas to limit the economic impact of the pan-demic.
3.4 THE IMPACT OF THE COVID-19 PANDEMIC ON FINANCIAL MARkETS
Finally, a few recent papers explore how theCOVID-19 pandemic is affecting financialmarket participants’ behaviour and percep-tions. Baker et al. (2020a) empirically docu-ment that no previous infectious disease out-break, including the Spanish Flu, impacted thestock market as strongly as the COVID-19pandemic. They attribute this to: (i) the easeand speed with which information on the pan-demic is disseminated, which generates highstock market volatility; (ii) the high intercon-nectedness of the world economy, whichimplies that economic disruption in one loca-tion has large spillover effects; but mostly to(iii) the COVID-19 containment policies,
which are much more extensive and wide-spread than similar efforts in the past and leadto a sharp decline in labour supply.
Hassan et al. (2020) construct time-varyingmeasures of the exposure of individual firms toCOVID-19, as well as measures of firm-spe-cific sentiment and riskiness with regard to theCOVID-19 pandemic. They do so for a globalsample of firms, by applying a text-classifica-tion technique on their quarterly earnings con-ference calls with market participants. Thesemeasures reflect both firms’ and markets’ con-cerns as, during the conference calls, firm man-agers have to respond directly to questionsfrom market participants about their firm’sprospects and thus address issues they mightnot have raised voluntarily. The authors thusidentify which firms are expected to gain orlose from the pandemic and which are mostaffected by the associated uncertainty. Theyfind that, in the first quarter of 2020, firms’ pri-mary concerns related to the collapse ofdemand, increased uncertainty, and disrup-tions in supply chains. Other important con-cerns relate to capacity reductions, closures,and employee welfare. By contrast, financingconcerns are mentioned relatively rarely. Alimited number of firms foresee opportunitiesin new or disrupted markets due to the spreadof the disease. Finally, there is some evidencethat firms which have experience with SARS orH1N1 have more positive expectations abouttheir ability to deal with the COVID-19 out-break.
Ramelli and Wagner (2020) focus on theindustry-level cross-section of returns and findthat, within the same industry and controllingfor standard firm characteristics, more lever-aged firms and those with limited cash holdingssuffered more severely, even those with littleor no international activities or linkages toChina. In contrast to Hassan et al. (2020), they
51Economic Bulletin
July 2020 29
21 In a similar vein, Briscese et al. (2020) show that ensuring thepublic’s expectations of the lockdown duration are unbiasedmatters for the success of the policy. Other papers draw onprinciples of behavioural economics to explore how COVID-19-related messages can be conveyed more effectively to the broaderpublic; see for example Haushofer and Metcalf (2020).
conclude that investors were mainly concernedabout firms with corporate debt and limitedliquidity, thus amplifying the COVID-19 eco-nomic crisis through financial channels.
Zechner et al. (2020) study how dividends havebehaved during the recent period of turbu-lence. They find that, although firms normallyattach great importance to smoothing their div-idend payouts, so as to provide shareholderswith projectable income streams, the oppositeis true in disaster states. Despite robust 2019earnings, many companies have slashed pre-viously announced dividends to protect theirliquidity, or are expected by the market to doso, while in some European countries regula-tors have forced companies to stop paying div-idends or have tied government subsidiesintended to help them cope with the crisis todividend cuts. Thus, it seems that the liquiditywhich dividends represent for shareholders dis-appears in precisely those states in which pre-dictable cash payments would be valued mosthighly. This explains the recent sharp increasein the risk premium on dividend claims.
3.5 THE IMPACT OF THE COVID-19 PANDEMIC ONLABOUR MARkETS AND DIFFERENT SEGMENTSOF SOCIETy
Several papers explore the ways in which theCOVID-19 pandemic has affected labour mar-kets and different social groups. Dingel andNeiman (2020) classify all occupations in termsof work-from-home feasibility. By merging thisclassification with occupational employmentdata for the US, they find that 37% of US jobscan plausibly be performed at home. Hensvicket al. (2020) compute the share of teleworkingundertaken in the US over the period 2011-18,by occupation and industry, as an indication ofthe extent to which teleworking could beemployed during the pandemic, and find thatit varies greatly.
Alon et al. (2020) explore whether the eco-nomic downturn caused by the currentCOVID-19 outbreak may have implications forgender equality during both the downturn and
the subsequent recovery. During typical reces-sions, male employment is affected moreseverely than female employment. Thisreflects both differences in the sectoral com-position of their employment (women tend tobe employed in more “secure” sectors, e.g. thegovernment) and the fact that women often optto increase their labour force participation asa response to their male partners’ employmentuncertainty. Conversely, the decline inemployment which stems from social distanc-ing measures may have a relatively largerimpact on sectors with high female employ-ment shares. In addition, school closuresincrease child care needs, which likely has aparticularly large impact on working mothers.The effects of the crisis on working mothersare likely to be persistent, due to high returnsto experience in the labour market. In the longrun however, the adoption of flexible workarrangements may ultimately promote genderequality in the labour market. Moreover, thefact that fathers may also often be obliged totake primary responsibility for child care dur-ing the epidemic could help erode discrimi-nating social norms and have a permanent pos-itive effect on male participation in child care,as is known to be the case for compulsorypaternity leave.
Allcott et al. (2020b) use location data from alarge sample of smartphones to show that, con-trolling for other factors, areas with moreRepublicans engage in less social distancing.They then present new survey evidence of sig-nificant gaps between Republicans andDemocrats in beliefs about their personal riskand the future path of the pandemic.
Baker et al. (2020b) study transaction-levelhousehold data and find that households dras-tically altered their spending behaviour, as thenumber of COVID-19 infections began togrow. Initially, spending increased sharply, par-ticularly in retail, credit card spending andfood items. This was followed by a sharpdecrease in overall spending. They detect sub-stantial heterogeneity across partisan affilia-tion, demographics and income. Specifically,
51Economic BulletinJuly 202030
Republicans were more likely to stockpile, pos-sibly because they are on average older, butalso perhaps because they are more concernedabout the financial implications of the epi-demic. Other cohorts that undertook relativelymore stockpiling were pensioners, householdswith children, and women.
4 ASyMMETRIES AND SPILLOVER EFFECTS OF COVID-19: kEy TAkEAWAyS FROM THE ASSESSMENT OF INTERNATIONALINSTITUTIONS
International organisations, such as theEuropean Commission, the OECD, the IMFand the World Bank, have attempted to quan-tify the effects of COVID-19 and to integrateits impact into their fully fledged forecasts forthe world economy. The quantification of theimpact, which is largely based on simulationanalysis, is surrounded by a high degree ofuncertainty, due to the unpredictability of fac-tors, such as the success of containment meas-ures and the possible occurrence of successiveoutbreaks. This section focuses on some oftheir findings with respect to the asymmetricimpact and the spillover effects of the COVID-19 pandemic within and across countries andregions.
The economic impact of COVID-19 can begreater in certain regions compared with oth-ers, mainly due to differences in financial con-ditions and available policy space. Emergingmarket economies (EMEs) are particularlyvulnerable to the financial channel of trans-mission of the crisis, compared with mostadvanced economies (AEs). This has been evi-dent in the tightening of financial conditionsin these countries and the unprecedented cap-ital outflows due to increased risk aversion anda flight by investors to safety and liquidity (see,for example, UNIDO 2020). Whereas a groupof 25 EMEs including China, India, SouthAfrica and Brazil had a net inflow of invest-ments of USD 79 billion in 2019, a total ofUSD 97 billion in portfolio equity and debtinvestments has already exited these countries
during 2020Q1, according to the Institute ofInternational Finance (2020). These capitalflows are larger than during any recent crisisepisode, including the global financial crisis in2008. The capital flight has renewed fears ofinsolvency and sovereign default. This could befurther accelerated by currency depreciationsin EMEs and notably in countries such asArgentina, Turkey or South Africa.
In addition, in many EMEs, fiscal automaticstabilisers are weaker relative to AEs, due tothe magnitude of the informal sector and lessdeveloped social safety nets. Under these con-ditions, macroeconomic policies to supportemployment and incomes, such as unemploy-ment benefits and subsidised leaves, have lim-ited effect. As a result, lockdown measures canbe more costly and lead to widespread unem-ployment and bankruptcies, with significantincome losses in lower-income economies, par-ticularly affecting the poorest members of soci-ety. ILO (2020) estimates for the impact ofCOVID-19 on global employment suggest thatlower-income countries are more vulnerable,largely due to higher informal employment.22
Moreover, although public and corporate sec-tors are highly leveraged across the world, theyare a particular source of vulnerability forEMEs. Coupled with the high share of exter-nal debt, debt denominated in foreign curren-cies and heavy reliance on short-term debt,EMEs could be subject to serious balancesheet mismatches. Several EMEs are also netenergy exporters and, hence, would be heavilyexposed to the negative supply shock of lowcommodity prices, which is mainly driven byplunging global energy demand.
The IMF (2020), in its World Economic Out-look of 13 April 2020, highlights these poten-tial asymmetric effects of COVID-19 betweenAEs and EMEs, notably due to tighter creditconditions and differences in policy space. The
51Economic Bulletin
July 2020 31
22 Lower middle-income countries are set to register the highest rateof working hours lost (12.5%) in 2020Q2, compared with the pre-crisis baseline (2019Q4). By contrast, working hours in high-incomecountries are set to decline by 11.6%.
asymmetries are manifested in both the short-term and the longer-term (“scarring”) effectsof the pandemic and in the containment meas-ures. More specifically, based on a globalmodel and a detailed sector-based analysis, theIMF presents a simulation exercise with threeadverse scenarios for global GDP comparedwith its forecast baseline, according to whichglobal GDP growth is projected at -3.0% for2020, i.e. a 6.3 percentage point downwardrevision compared with the January 2020 WEOUpdate. The three scenarios assume: (i) a lock-down that lasts 50% longer than in the base-line; (ii) a second outbreak in 2021; and (iii)both a longer lockdown and a second outbreak,respectively. The model assumes that theimpact is driven by containment measures anda tightening of financial conditions, and is mit-igated by fiscal and monetary policy measures.Despite policy support, the pandemic leavesscarring effects on capital, productivity andtrend employment.
Moreover, the model incorporates asymmetriceffects of COVID-19 brought about by the lackof available policy space in EMEs, which lim-its their ability to improve financial marketconditions and to mitigate the scarring on theeconomy, compared with AEs. As a result, theimpact of COVID-19 on this group of coun-tries is amplified. The decline in GDP for 2020and 2021 is estimated to be of similar magni-tude to that in AEs, despite the fact that theservices sectors, which are most affected byCOVID-19, have a relatively smaller economicsignificance in EMEs compared with AEs.Notably, output decline in the medium term isgreater in EMEs relative to AEs due to theineffectiveness of policy in mitigating the scar-ring of the economy. In the most adverse sce-nario, the output loss in 2024 is around 3.5%for AEs, against almost 4.5% for EMEs.
The World Bank (2020a), in a publication on8 April 2020, presents a simulation exerciseusing a BVAR model to predict growth fordeveloping countries in the “Europe and Cen-tral Asia” (ECA) region, compared with theJanuary 2020 projections. Trade, transport and
tourism, as well as financial conditions andcommodity prices are assumed to be the maintransmission channels of the COVID-19impact for this particular group of countries.Global financial stress and flight-to-safety putpressure on currencies and are expected tolead to tightened financial conditions, withpossible negative repercussions for corporatebalance sheets. Also, given that countries in theregion are significant energy exporters, the fallin oil and metal prices, mainly due to reducedimports from China, is expected to affectexports and strain fiscal positions in the region.The results of the analysis predict 5.4 per-centage points lower GDP growth for theregion in 2020 in the baseline scenario, and 7.0percentage points lower growth in the down-side scenario, which assumes that the con-tainment measures, financial market pressuresand low commodity prices last beyond2020H1. This translates into a GDP growthrate of -2.8% for 2020 in the baseline scenarioand of -4.4% in the downside scenario.
Kohlscheen et al. (2020), in a BIS paper of 6April 2020, simulate the propagation ofCOVID-19 putting emphasis on the amplifica-tion effects from spillovers across countries.23 Inparticular, they employ a quarterly globalBVAR over the period 1997-2019 with fivemajor economic blocs: the US, China, the euroarea, “other advanced economies” and “otherEMEs”. The economic impact of the virusdepends on: (i) the direct effects of confine-ment measures and their duration; (ii) theextent to which the direct effects persist andmagnify; and (iii) the size of spillovers and spill-backs across regions. The model sheds light onthe multiplier effects of the initial slowdown inactivity, the persistence of the slowdown withineach region, and the extent of spillovers.
Global economic spillovers are set to be large.The authors estimate that, on average, the full-year GDP loss in the regions included in themodel would be between 1.5 and 2 times the ini-
51Economic BulletinJuly 202032
23 The group of “other advanced economies” comprise Australia,Canada, Switzerland, the UK, Japan and Sweden, while the groupof “other EMEs” comprise Brazil, India, South Korea and Mexico.
tial impulse from containment measures. Intheir V-shaped scenarios, the recovery in2020H2 is modest, and even at end-2021, thelevel of GDP in all regions would still be belowthe pre-virus forecast. In their W-shaped sce-narios (i.e. a second wave of confinement fol-lows two quarters after the first wave), the weak-ness in economic activity persists for evenlonger; in most regions examined, GDP growthis negative throughout 2020 and a sustainedrecovery would not begin until 2021, or aroundsix months later than in the V-shaped scenarios.
The persistence of weak activity partlyreflects two types of spillovers. One is due tothe risk that uncoordinated lockdowns lead torepeated virus outbreaks and confinementsacross the globe. The other is the more tradi-tional trade and financial interlinkages. ForAEs, spillovers from EMEs account forbetween 25%-30% of the GDP shortfall in2020Q4. The spillovers are larger for the euroarea, due to a larger share of exports in GDP,than for the US. Moreover, domestic mitiga-tion alone is ineffective. Even if a country suc-cessfully limits its domestic slowdown, it willnot be immune from insufficient or ineffectivepolicies in other parts of the world. The authorsrun a scenario where the effect of domesticcontainment measures on euro area GDP is -2.5%, but the shock still hits other regions by-5%. The decline in euro area GDP after fourquarters would still be 6.5% in the V-shapedand 9.9% in the W-shaped scenario, relative tothe baseline. This outcome reinforces theimportance of international cooperation indesigning policies to limit the spread and there-emergence of the virus and combat its eco-nomic consequences.
In the same vein, the OECD’s (2020a) esti-mates in early March 2020, based on NiGEMsimulations for the G20, highlight the positivespillover effects induced from policy coordi-nation.24 These suggest that collective countryaction yields higher output gains than individ-ual country responses via positive confidenceand trade spillovers; in particular, coordinatedfiscal, monetary and structural policies should
raise the level of GDP by 0.75% in the firstyear, 1.25% in the second year and 1.0% in thelong run. By contrast, individual countryresponses would increase GDP by only 0.4%,0.75% and 0.7%, respectively.
Moreover, several international institutionshighlight the fact that the asymmetric impactof COVID-19 is driven by a varying sectoraldemand composition, reflecting the share ofthe affected sectors in consumption spendingand total output.
More specifically, in a two-step approach, theOECD (2020b) in late March 2020 (updated inmid-April) provides an illustrative exercise onthe initial, short-term impact of the lockdownfrom COVID-19 on the level of real GDP inOECD economies. First, from a sectoral out-put approach, it is assumed that value addeddeclines by 50%-100% in the sectors affectedby the lockdown. Second, from a spendingapproach, cutbacks in categories of consumerspending are assumed to range between 50%and 100%. Common effects are assumed withinsectors in all countries. A caveat is that cross-sectoral spillovers, potential indirect effects orother offsetting factors, such as policy meas-ures, are not taken on board. Estimates suggestthat real output loss on impact ranges between20% and 25% in the G7, and between 15% and35% in major AEs and EMEs, the highestbeing in Greece. Differences across countriesreflect a varying sectoral composition of out-put. The spending-based assessment underpinsan even sharper short-term impact on con-sumer spending in all countries.
The European Commission (2020) ―in addi-tion to its DSGE simulations for the impact ofCOVID-19 on the EU economy presented inits Spring 2020 Economic Forecast on 6 May2020― puts forward an input-output sectoralmodel to assess the sectoral and countryspillovers from lockdown measures during the
51Economic Bulletin
July 2020 33
24 Coordinated policy responses include a debt-financed fiscal easingof 0.5% of GDP in all countries for three years, a reduction ininterest rates and competition-enhancing reforms which raise TFPby 1% after five years.
COVID-19 pandemic. Based on the Trade-SCAN input-output multi-country model,simultaneous shocks to sectoral final demandare applied to the EU and the rest of the world.Final demand is assumed to decline by about5%, which is consistent with the respectiveQUEST “baseline” simulations. Model esti-mates imply a high degree of propagation ofdemand shocks across countries and sectors,with the final effect being higher than the ini-tial direct hit to demand; euro area GDP con-tracts by 5.7% in 2020 on average, while out-put losses range between -5.0% in Finland and-8.0% in Greece and Malta. Differences acrossMember States reflect their relative exposureto tourism and the importance of input-outputspillovers in the tourism sector.
Finally, the WTO (on 8 April 2020) simulatesthe GDP effects of the COVID-19 pandemicusing a recursive dynamic CGE model, namelythe WTO Global Trade Model, which containsdetailed sectoral breakdowns and intermediatelinkages that enable the study of upstream anddownstream effects of the sectoral shocks. TheWTO presents three alternative scenarios. Inthe optimistic scenario, the containment meas-ures will stay in place for three months, fol-lowed by a V-shaped recovery. In the less opti-mistic scenario, measures stay in place for sixmonths, leading to a U-shaped recovery. In thepessimistic scenario, the measures will have tostay in place for the entire 2020, leading to anL-shaped recovery, as heightened economicuncertainty postpones consumption of durablemanufactured goods. The short-term globaloutput losses range between 5% and 11% in2020 relative to the baseline, while regionalpatterns show the biggest output drops inEMEs. ASEAN (South-East Asia), Mexico andthe Newly Industrialised Countries (e.g. Korea,Hong Kong, Taiwan) are projected to see thesharpest decline in GDP. For the US, thereduction in trade is projected to be muchlarger than the reduction in GDP, as the shareof goods traded by air and the share of exportsin services are large. Moreover, the relativecontribution of different shocks changes overtime. In the V-shaped scenario, labour supply,
trade costs and sectoral demand shocks con-tribute 42%, 20%, and 38%, respectively, tothe fall in global GDP. In the U-shaped and L-shaped scenarios, the contribution of thesectoral demand shocks rises to above 50%.The largest differences among countries aredriven by the sectoral demand component,reflecting the high share of these sectors intotal household consumption.
5 A BIRD’S EyE VIEW: CHANNELS, IMPACT AND POLICy IMPLICATIONS
This comprehensive overview of variousstrands of the literature and of empiricalassessments by international institutionsallows us to draw some conclusions about theeconomic impact of pandemics in general andof COVID-19 in particular.
First, we can identify the channels throughwhich pandemics affect the economy. Themost obvious channel comes from supply-sideeffects. Pandemics reduce both the quantityand the quality of labour. Crucially, the quan-titative impact will depend on the mortalityrate of the epidemic, as well as on the extentto which the working age population isaffected. Labour supply may also be affectedby the exclusion of different social groups whoare deemed likely to spur the spreading of thepandemic, as well as through the impact of thepandemic on migratory flows. In the shortterm, the labour supply effect is also influ-enced by the morbidity rate – the extent towhich lockdowns lead to workers being unableto work during that period. In losing parts ofthe working age population, human capitaleffects are also likely to influence the qualityof labour supply in the post-pandemic era.These effects, working through labour supply,negatively impact on both actual and poten-tial output, pointing to longer-term effectsfrom pandemics. Pandemics can also generatedestruction of capital effects, as businessesclose and investment is curtailed, thus addingto the negative impact on current and poten-tial output.
51Economic BulletinJuly 202034
Aside from supply-side effects, pandemics arelikely to induce demand effects. Consumptionis particularly vulnerable to the impact of bothreduced income and increasing uncertaintywhich dents consumers’ confidence. Uncer-tainty will also have the effect of, at best, delay-ing investment and, at worst, dampening it fora period of time. Overall, the savings rate inthe private sector should increase as uncer-tainty over a future health care crisis looms andcash flow constraints pose challenges to debtservicing.
Moreover, the impact of both supply anddemand effects is also likely to depend on thereaction of individuals. How quickly and towhat extent workers lock down is determinedby their beliefs about their exposure to thevirus and their views regarding the ability ofthe health system to cope. Individual reactionswill also impact on the ability of the economyto quickly exit lockdown. Even if statutorymeasures are relaxed, individual behaviourmay take more time to adjust.
A final channel works through the financialsystem. While the natural rate of interest mightbe expected to fall, leading to a period of lowinterest rates, financial institutions are likelyto come under increasing stress. Rising uncer-tainty, along with an increase in the number ofborrowers with debt servicing difficulties, islikely to generate a liquidity squeeze, whichexacerbates the demand effects of pandemics.Moreover, risk premia are expected to rise andthe evidence suggests that COVID-19 has hadstronger effects on financial markets than pastpandemics.
Whether it is the supply channel or the demandchannel that dominates is not so important.Even if a shock originates from the demandside, it can then feed back into the supply sidethrough the investment channel. Lowerdemand reduces labour productivity, as outputfalls. This then generates lower investment,which further depresses labour productivityand, ultimately, demand. This is what Fornaroand Wolf (2020) call the “supply-demand
doom loop”. Such doom loops are common inmodels that include more than one sector –one particularly affected by the lockdown(high-contact sectors) and a second which isless affected. The shock to the most affectedsector is easily passed on to less affected sec-tors, creating vicious circles.
Building upon these channels it is possible toidentify the impact of pandemics on some keyeconomic aggregates. All three channels workto reduce current and possibly potential output.Growth is negatively affected in the short run,but usually rebounds in subsequent years. How-ever, output can still remain below trend. Theimpact on inflation depends on the relative bal-ance of supply versus demand shocks. If thesupply shocks are large, as they were during theBlack Death, then inflation can rise as short-ages develop. On the other hand, if demandeffects dominate, deflationary pressures will bepresent. With respect to inflation, the impactof the demand shock on commodity prices alsoworks to lower inflationary pressures. Theimpact on trade is usually greater than that onoutput and this makes open economies andeconomies that are closely embedded in globalvalue chains especially vulnerable. Travel isusually one of the most affected sectors andthus economies which rely on tourism exportswill be more negatively affected than those thatare primarily goods exporters. Goods exporters,in today’s world, aside from experiencing theimpact of weaker demand, also face potentialsupply shortages, as production chains are dis-rupted. The destruction of livelihoods, as busi-nesses close and unemployment rises, can havenegative implications for poverty, especially incountries that lack a strong welfare state. Ingeneral, the distributional impact of the pan-demic and its particularly strong impact on spe-cific sectors suggest the need for governmentsto engage in redistributional policies in orderto share the burden more fairly across society.Finally, social capital can also be destroyed dur-ing pandemics, as curtailing social interactionleads to a decline in trust. Lower social capitalhas been shown to reduce growth prospectsover the longer term.
51Economic Bulletin
July 2020 35
The negative effects of the pandemics can beasymmetric across sectors and regions due tovarious factors. First, uncoordinated confine-ments can lead to repeated and unsynchro-nised virus outbreaks across the globe. Second,the absorptive capacity of the exogenous healthshock depends on country-specific idiosyn-cratic factors, such as labour market flexibility,foreign capital dependence, trade opennessand policy space. Differences in the initial eco-nomic conditions can lead to a varying impactof the pandemic across countries. Third, tradeand financial linkages, among others, increasespillovers and amplify the effects of the first-order demand and supply shocks from the pan-demic, notably from COVID-19. The totalGDP shortfall from COVID-19 could be asmuch as twice the direct impact of the virusand the confinement measures, highlightingthe sizeable effect of multipliers and spilloversin propagating contractions within and acrosseconomies. These spillovers are largely evidentbetween AEs and EMEs, but also among AEs.
The negative effects of the pandemics can, ofcourse, be mitigated by policy, and the policytools that governments have at their disposalare much broader than those that were avail-able to countries during the two largest pan-demics – the Black Death and the Spanish Flu.The channels outlined above provide the back-ground for discussing policy responses.
The supply-side effects: One of the strongesteffects from past pandemics from the supplyside arises from the loss of life and hence theimpact on labour supply. COVID-19 is notexpected to have such a direct effect on laboursupply. The lockdowns have limited the impactof the pandemic on health, and labour supplyis disrupted in the short term more or lessdepending on the length of the lockdown. Thesupply-side effects in this pandemic are likelyto arise from the destruction of economic rela-tions – that is, the laying off of workers and thesubsequent rise in unemployment along withbusiness failures, which lead to the destructionof capital. Policy is thus focused on minimisingthe scarring effects by maintaining economic
relations as intact as possible through the lock-downs. Thus, policies to subsidise wages helpprevent workers from being laid off. The morewidespread use of wage subsidies in Europecompared to the US helps to explain the dif-fering unemployment outcomes. Similarly,policies to provide liquidity support to busi-nesses aim at keeping viable firms alive. In thisway, when lockdowns end, economic relationsare still fairly intact and it is easier to get pro-duction up and running. Of course, some busi-nesses will fail and unemployment will rise. Toprevent the latter from having lasting effectson the quality of labour supply, Active LabourMarket Policies (ALMPs) have to be a prior-ity area for improvement, especially in Greecewhere their past performance has been some-what patchy.
The demand-side effects: Monetary policy loos-ening either through interest rate reductions orthe extension of non-standard measures canhelp ease liquidity constraints for companies bypreventing liquidity shortages either directly orindirectly (through banks or various financialmarkets which have seized up). The ECB’sPandemic Emergency Purchase Programme(PEPP) is a step in the right direction to ensur-ing that liquidity provision is maintained.
However, fiscal policy also has to play a sig-nificant role. Most countries have beenannouncing large fiscal stimuli whether it isthrough transfers to subsidise wages, measuresto delay tax and social security payments, orthe extension of eligibility for unemploymentbenefits beyond their usual fixed period. Con-siderations of redistribution across more andless affected groups are also beneficial inencouraging recovery.
Aside from financing transfers, fiscal policyalso has to support investment ―both privateand public― in the recovery period. The largeincrease in uncertainty will have led to thepostponement of investment plans and makesbusiness less optimistic in developing newplans. As an individual business, it is rationalto postpone. An investment will only be prof-
51Economic BulletinJuly 202036
itable if others also invest raising aggregatedemand, productivity and hence disposableincome and, finally, consumption. The publicsector has a role to play in this regard, sincesuch investments usually generate large mul-tiplier effects in the economy and are likely tomake smaller private sector investments moreprofitable. Thus, EU policies for injectingfunds into investment projects ―eitherdirectly through Commission resources or viathe European Investment Bank― are wel-come. Moreover, investment raises productiv-ity, which can contribute to a demand reboundand help avoid the demand-supply doom loopsfound in the literature.
However, such stimuli will place an increasingburden on debt levels throughout the euroarea. For this reason, some form of mutuali-sation of the cost is necessary if national pro-grammes are to be feasible and credible. Somehave argued that the EU should issue perpet-ual bonds (Giavazzi and Tabellini 2020; Soros2020). This proposal is a sound one ―the EUhas the economic and political power to issuesuch bonds just as the UK and the US havedone in the past (the UK, for example, usedperpetual bonds to finance both theNapoleonic Wars and WWI). The EU also hasthe motive― a powerful symmetric shock fac-ing individual Member States. Indeed, the veryact of issuing such bonds would send a strongsignal to markets about government commit-ments to the EU project.
Financial channel: we can also draw some pol-icy lessons for the financial sector. The finan-cial sector has the potential to exacerbate thesupply and demand shocks that are the resultof a pandemic. First, financial institutions arelikely to face liquidity constraints, as interbankmarkets dry up and, potentially, deposits fall.Second, the impact of the pandemic on incomeand employment causes debt servicing diffi-culties among both households and firms. Thisin turn is likely to lead to rising non-perform-ing loans (NPLs), at a time when a number ofEU countries already face higher than usualNPL ratios.
Potential liquidity constraints in euro areabanks are being addressed by the Eurosystem.First, Greek government bonds (GGBs) havebecome eligible both for the PEPP and also ascollateral. Moreover, haircuts associated withmonetary transactions on all collateral (includ-ing GGBs) have been reduced significantly.The ability to tap liquidity from the Eurosys-tem using credit claims has been expanded. Allthese measures improve the liquidity situationof Eurosystem banks in general and Greekbanks in particular, as they provide the oppor-tunity for banks to resort to the Eurosystem forliquidity in greater amounts rather than rely-ing on interbank markets, which often prove tobe more expensive and susceptible to suddenshifts in sentiment.
Turning to the impact of the crisis on the assetquality of banks, evidence suggests that assetquality will deteriorate. At a time when NPLs arealready high in certain EU countries, this cir-cumstance poses a particular challenge. Nationalsystems to deal with NPLs may not be enough,since they rely on investors from elsewhere toinvest in the NPL clean-up. If, however, thedemand for such capital injections from outsideinvestors rises, it is not clear that there will beenough outside investors willing to meet thatdemand. The more systemic the increase inNPLs and thus the need to find a solution acrossthe euro area/EU as a whole, the more likely itis that some form of centralised solution mightbe required. It is against this background thatcalls for an EU-wide bad bank are being heard.
6 CONCLUSIONS
In short, pandemics have strong impacts oneconomies – both in the short run and poten-tially in the long run. Research relying on his-torical episodes as well as modern modellingtechniques can shed light on the channelsthrough which economies are affected. How-ever, modern states have many more policyweapons at their disposal and the models pro-vide ample evidence on what can be done tolimit the impact of such pandemics. Many of
51Economic Bulletin
July 2020 37
these policies are already being employed. Inthe context of the EU, their credibility wouldbenefit considerably from a more coordinatedand mutualised response to the crisis.
Finally, we can highlight the need for globalcooperation. Although policy efforts can miti-gate the adverse direct effects of the exogenousshock and preserve economic relationships,incomes and production structures, differencesin the available policy space can magnify struc-tural divergences across economies. Advancedeconomies are expected to be more effective
than emerging market economies in copingwith the effects of the pandemic as a result oflarger policy space. Still, divergences can beobserved even among advanced economies. Inthis regard, domestic mitigation alone is notsufficient to cope with the crisis. Even if coun-tries successfully limit the domestic economicslowdown, they will not be immune from insuf-ficient or ineffective policies in other countries.Hence, positive spillovers stemming from pol-icy coordination can underpin domesticefforts to lessen the economic hardship acrossthe world as a whole.
51Economic BulletinJuly 202038
Allcott, Η., L. Boxell, J.C. Conway, M. Gentzkow, M. Thaler and D.Y. Yang (2020), “ Polarizationand public health: partisan differences in social distancing during the coronavirus pandemic”,NBER Working Paper No. 26946.
Almond, D. (2006), “Is the 1918 influenza pandemic over? Long-term effects of in utero influenzaexposure in the post-1940 US population”, Journal of Political Economy, 114(4), 672-712.
Alon, T.M., M. Doepke, J. Olmstead-Rumsey and M. Tertilt (2020), “The impact of COVID-19 on gender equality”, NBER Working Paper No. 26947.
Atkeson, A. (2020), “What will be the economic impact of COVID-19 in the US? Rough esti-mates of disease scenarios”, NBER Working Paper No. 26867.
Baker, S.R., N. Bloom, S.J. Davis, K. Kost, M. Sammon and T. Viratyosin (2020a), “The unprece-dented stock market reaction to COVID-19”, Becker Friedman Institute White Paper, March.
Baker, S.R., R.A. Farrokhnia, S. Meyer, M. Pagel and C. Yannelis (2020b), “How does house-hold spending respond to an epidemic? Consumption during the 2020 COVID-19 pandemic”,NBER Working Paper No. 26949.
Baldwin, R. and B.W. di Mauro (eds.) (2020a), Economics in the Time of COVID-19, AVoxEU.org Book, CEPR Press.
Baldwin, R. and B.W. di Mauro (eds.) (2020b), Mitigating the COVID Economic Crisis: Act Fastand Do Whatever It Takes, A VoxEU.org Book, CEPR Press.
Barro, R.J. (2020), “Non-pharmaceutical interventions and mortality in U.S. cities during theGreat Influenza Pandemic, 1918-1919”, NBER Working Paper No. 27049.
Barro, R.J., J.F. Ursúa and J. Weng (2020), “The coronavirus and the Great Influenza Pandemic:lessons from the ‘Spanish Flu’ for the coronavirus’s potential effects on mortality and economicactivity”, NBER Working Paper. No. 26866.
Bartik, A.W., M. Bertrand, Z.B. Cullen, E.L. Glaeser, M. Luca and C.T. Stanton (2020), “Howare small businesses adjusting to COVID-19? Early evidence from a survey”, NBER WorkingPaper No. 26989.
Bayer, C., B. Born, R. Luetticke and G. Müller (2020), “The coronavirus stimulus package: howlarge is the transfer multiplier?”, CEPR Discussion Paper No. DP14600.
Berger, D.W., K.F. Herkenhoff and S. Mongey (2020), “An SEIR infectious disease model withtesting and conditional quarantine”, NBER Working Paper No. 26901.
Bloom, D. and A. Mahal (1997), “AIDS, flu, and the Black Death: impacts on economic growthand well-being”, The Economics of HIV and AIDS: The Case of South and South East Asia,Delhi: Oxford University Press, 22-52.
Bloom, E., V. de Wit and M.J. Carangal-San Jose (2005), “Potential economic impact of an avianflu pandemic on Asia”, ERD Policy Brief No. 42, Asian Development Bank, Economics andResearch Department, Manila.
Bootsma, M.C. and N.M. Ferguson (2007), “The effect of public health measures on the 1918influenza pandemic in US cities”, Proceedings of the National Academy of Sciences, 104(18),7588-7593.
Brainerd, E. and M.V. Siegler (2003), “The economic effects of the 1918 influenza epidemic”,CEPR Discussion Paper No. DP3791.
Briscese, G., N. Lacetera, M. Macis and M. Tonin (2020), “Compliance with COVID-19 socialdistancing measures in Italy: the role of expectations and duration”, NBER Working Paper No.26916.
Burns, A., D. van der Mensbrugghe and H. Timmer (2006), “Evaluating the economic conse-quences of avian influenza”, World Bank Working Paper No. 47417.
Casares, M. and H. Khan (2020), “A dynamic model of COVID-19: contagion and implicationsof isolation enforcement”, Carleton Economics Working Paper No. CEWP 20-02, Canada.
51Economic Bulletin
July 2020 39
R E F E R ENC E S
Chou, J., N. Kuo and S. Peng (2003), “The potential impacts on the Taiwanese economy of theoutbreak of SARS”, Paper presented at Asian Economic Panel, Keio University, Tokyo, 11-12 May.
Cirillo, P. and N.N. Taleb (2020), “Tail risk of contagious diseases”, Nature Physics, 16Correia, S., S. Luck and E. Verner (2020), “Pandemics depress the economy, public health inter-
ventions do not: evidence from the 1918 flu”, Mimeo.Dingel, J. and B. Neiman (2020), “How many jobs can be done at home?”, CEPR, DiscussionPaper Series, DP14584.
Eichenbaum, M., S. Rebelo and M. Trabandt (2020), “The macroeconomics of epidemics”, CEPRDiscussion Paper No. DP14520.
European Commission (EC) (2020), European Economic Forecast, Spring 2020, 6 May. Fan, E.X. (2003), “SARS: economic impacts and implications”, ERD Policy Brief No. 15, Asian
Development Bank, Economics and Research Department, Manila.Fan, V.Y., D.T. Jamison and L.H. Summers (2016), “The inclusive cost of pandemic influenza
risk”, NBER Working Paper No. 22137.Faria-e-Castro, M. (2020), “Fiscal policy during a pandemic”, Federal Reserve Bank of St. Louis
Working Paper No. WP 2020-006A.Ferguson, N.M, D. Laydon, G. Nedjati-Gilani, N. Imai, K. Ainslie, M. Baguelin, S. Bhatia, A.
Boonyasiri, Z. Cucunubá, G. Cuomo-Dannenburg, A., Dighe, I. Dorigatti, H. Fu, K. Gaythorpe,W. Green, A. Hamlet, W. Hinsley, L.C. Okell, S. van Elsland, H. Thompson, R. Verity, E. Volz,H. Wang, Y. Wang, P.G.T. Walker, C. Walters, P. Winskill, C. Whittaker, C.A. Donnelly, S. Rileyand A.C. Ghani (2020), “Report 9: Impact of non-pharmaceutical interventions (NPIs) to reduceCOVID19 mortality and healthcare demand”, Imperial College COVID-19 Response Team.
Fernandez, Nuno (2020) “Economic effects of coronavirus outbreak (COVID-19) on the worldeconomy”, mimeo.
Fetzer T., L. Hensel, J. Hermle and C. Roth (2020), “Coronavirus perceptions and economic anx-iety”, Cornell University arxiv.org, https://arxiv.org/pdf/2003.03848.pdf.
Fornaro, L. and M. Wolf (2020), “COVID-19 coronavirus and macroeconomic policy: some ana-lytical notes,” CREI/UPF and University of Vienna.
Garrett, T.A. (2009), “War and pestilence as labor market shocks: US manufacturing wage growth1914-1919”, Economic Inquiry, 47(4), 711-725.
Giavazzi, F. and G. Tabellini (2020), “COVID perpetual eurobonds: jointly guaranteed and sup-ported by the ECB”, VoxEU.org, March, https://voxeu.org/article/covid-perpetual-eurobonds.
Glover, A., J. Heathcote, D. Krueger and J.-V. Rios-Rull (2020), “Health versus wealth: on thedistributional effects of controlling a pandemic”, CEPR Discussion Paper No. DP14606.
Gourinchas, P.O. (2020), “Flattening the pandemic and recession curves”, Mitigating the COVIDEconomic Crisis: Act Fast and Do Whatever It Takes, A VoxEU.org eBook, CEPR Press, p. 31.
Greenstone, M. and V. Nigam (2020), “Does social distancing matter?”, Becker Friedman Insti-tute Working Paper No. 2020-26.
Guerrieri, V., G. Lorenzoni, L. Straub and I. Werning (2020), “Macroeconomic implications ofCOVID-19: can negative supply shocks cause demand shortages?”, NBER Working Paper No.26918.
Hassan, T., S. Hollander, L. van Lent and A. Tahoun (2020), “Firm-level exposure to epidemicdiseases: COVID-19, SARS, and H1N1”, CEPR Discussion Paper No. DP14573.
Hatchett, R.J., C.E. Mecher and M. Lipsitch (2007), “Public health interventions and epidemicintensity during the 1918 influenza pandemic”, Proceedings of the National Academy of Sci-ences, 104(18), 7582-7587.
Haushofer, J and C.J.E. Metcalf, (2020), “Combining behavioral economics and infectious dis-ease epidemiology to mitigate the COVID-19 outbreak”, Princeton mimeo.
51Economic BulletinJuly 202040
Hensvik, L., T. Le Barbanchon and R. Rathelot (2020), “Which jobs are done from home? Evi-dence from the American Time Use Survey”, CEPR Discussion Paper No. 14611.Hirshleifer, J. (1987), Economic behaviour in adversity, University of Chicago Press. Holmström, B. and J. Tirole (1998), “Private and public supply of liquidity”, Journal of Politi-cal Economy, 106(1), 1-40.Institute of International Finance (2020), “Capital Flows Report”, 9 April, https://www.iif.com/Publications/ID/3841/Capital-Flows-Report-Sudden-Stop-in-Emerging-Markets.International Labour Organization (ILO) (2020), ILO Monitor 3nd edition: COVID-19 and theworld of work. Updated estimates and analysis, 29 April.IMF (2020), World Economic Outlook, International Monetary Fund, April.Jedwab, R., N.D. Johnson and M. Koyama (2019), “Pandemics, places, and populations: evidencefrom the Black Death”, CEPR Discussion Paper No. DP 13523.Jonas, O.B. (2013), Pandemic risk, World Bank, Washington DC.Jones, C.J., T. Philippon and V. Venkateswaran (2020), “Optimal mitigation policies in a pan-demic: social distancing and working from home”, NBER Working Paper No. 26984.Jordà, O., S.R. Singh and A.M. Taylor (2020), “Longer-run economic consequences of pan-demics”, Federal Reserve Bank of San Francisco Working Paper No. 2020-09, March.Karlsson, M., T. Nilsson and S. Pichler (2014), “The impact of the 1918 Spanish flu epidemicon economic performance in Sweden: an investigation into the consequences of an extraordi-nary mortality shock”, Journal of Health Economics, 36, 1-19. Kohlscheen, E., B. Mojon and D. Rees (2020), “The macroeconomic spillover effects of the pan-demic on the global economy”, BIS Bulletin, 4.Krueger, D., H. Uhlig and T. Xie (2020), “Macroeconomic dynamics and reallocation in an epi-demic”, CEPR Discussion Paper No. DP14607.Laeven, L. and F. Valencia (2020), “Systemic banking crises database: a timely update in COVID-19 times”, CEPR Discussion Paper No. DP14569.Lee, J.W. and W.J. McKibbin (2004), “Globalization and disease: the case of SARS”, Asian Eco-nomic Papers, 3(1), 113-131.Le Moglie, M., F. Gandolfi, G. Alfani and A. Aassve (2020), “Epidemics and trust: the case ofthe Spanish Flu”, IGIER Working Paper No. 661.Leiva-Leon, D., G. Pérez-Quirós and E. Rots (2020), “Real-time weakness of the global econ-omy: a first assessment of the coronavirus crisis”, CEPR Discussion Paper No. DP14484.Lilley, A., M. Lilley and G. Rinaldi (2020), “Pandemics, public health and economic growth: revis-iting the Spanish Flu evidence”, Harvard University mimeo.Loose, V.W., V.N. Vargas, D.E. Warren, S.J. Starks, T.J. Brown and B.J. Smith (2010), Economicand Policy Implications of Pandemic influenza, Report SAND2010-1910, Sandia National Lab-oratories, Albuquerque, New Mexico, March.Luo, S. and K.P. Tsang (2020), “How much of China and world GDP has the coronavirusreduced?”, mimeo, Department of Economics, Virginia Tech.Maddison, A. (2010), “Historical statistics on world population, GDP and per capita GDP, 1-2008 AD”, University of Groningen.Markel, H., A. Stern and M. Cetron (2008), “Nonpharmaceutical interventions implemented byUS cities during the 1918-1919 influenza pandemic”, International Journal of Infectious Dis-eases, 12, e432.McKibbin, W.J. and R. Fernando (2020), “The global macroeconomic impacts of COVID-19:seven scenarios”, Brookings Research, https://www.brookings.edu/research/the-global-macro-economic-impacts-of-covid-19-seven-scenarios/.McKibbin, W.J. and A. Sidorenko (2006), “Global macroeconomic consequences of pandemicinfluenza”, Sydney, Australia: Lowy Institute for International Policy.
51Economic Bulletin
July 2020 41
Meltzer, M.I., N.J. Cox and K. Fukuda (1999), “The economic impact of pandemic influenza inthe United States: priorities for intervention”, Emerging infectious diseases, 5(5), 659.Munro, J. (2005), “Before and after the Black Death: money, prices, and wages in fourteenth-century England”, University of Toronto, Department of Economics, Working Paper. OECD (2020a), Economic Outlook, Interim Report, March.OECD (2020b), “Evaluating the initial impact of COVID-19 containment measures on economicactivity”, March.Ramelli, S. and A.F. Wagner (2020), “Feverish stock price reactions to COVID-19”, CERP Dis-cussion Paper No. DP14511.Robbins, H. (1928), “A comparison of the effects of the Black Death on the economic organi-zation of France and England”, Journal of Political Economy, 36(4), 447-479.Simonsen, L., M.J. Clarke, L.B. Schonberger, N.H. Arden, N.J. Cox and K. Fukuda (1998), “Pan-demic versus epidemic influenza mortality: a pattern of changing age distribution”, Journal ofinfectious diseases, 178(1), 53-60.Siu, A. and Y.C.R. Wong (2003), “Ravaged by SARS: the case of Hong Kong SARS”, Paper pre-sented at Asian Economic Panel, Keio University, Tokyo, 11-12 May.Soros, G. (2020), “The EU should issue perpetual bonds”, Project Syndicate, https://www.pro-ject-syndicate.org/commentary/finance-european-union-recovery-with-perpetual-bonds-by-george-soros-2020-04.Stock, J.H. (2020), “Data gaps and the policy response to the novel coronavirus”, NBER Work-ing Paper No. 26902.Torój, A. (2013), “Why don’t Blanchard-Kahn ever ‘catch’ flu? And how it matters for measur-ing indirect cost of epidemics in DSGE framework”, Central European Journal of EconomicModelling and Econometrics, 5, 185-206. United Nations Industrial Development Organization (UNIDO) (2020), “Coronavirus: the eco-nomic impact”, 7 April.Velasco, A. and R. Chang (2020), “Economic policy incentives to preserve lives and livelihoods”,CEPR Discussion Paper No. DP14614.Velde, F.R. (2020), “What happened to the US economy during the 1918 influenza pandemic?A view through high-frequency data”, Federal Reserve Bank of Chicago Working Paper 2020-11.Voigtländer, N.o and H.-J. Voth (2012), “Persecution perpetuated: the medieval origins of anti-semitic violence in Nazi Germany,” Quarterly Journal of Economics, 127(3), 1-54. World Bank (2020a), Europe and Central Asia Economic Update, Spring 2020: Fighting COVID-19, World Bank, April.World Bank (2020b), East Asia and Pacific Economic Update, Spring 2020: East Asia and Pacificin the Time of COVID-19, April.World Trade Organization (WTO) (2020), “Methodology for the WTO trade forecast of April8 2020”.Zechner, J., G. Cejnek and O. Randl (2020), “The COVID-19 pandemic and corporate dividendpolicy”, CEPR Discussion Paper No. DP14571.
51Economic BulletinJuly 202042
Dimitris MalliaropulosEconomic Analysis and Research Departmentand University of Piraeus, Department of Banking and Financial Management
Petros MigiakisEconomic Analysis and Research Department
ABSTRACTWe discuss the factors behind sovereign credit ratings and reproduce their quantitative compo-nent, focusing on the case of Greece. The sovereign credit rating of Greece is still lower than theinvestment grade threshold. However, some of the fundamentals of the Greek economy are shownto be better than the average of the rating category it belongs to at present (BB) and better eventhan higher rating categories. Based on the reproduction of the score component of sovereign creditratings of the three major Credit Rating Agencies, we show that an improvement in the institu-tional factors of the Greek economy to the level of the early 2000s can lead to a significant increasein Greece’s score, thus contributing to an upgrade to the investment grade category.
Keywords: credit ratings; sovereign risk; Greek economy; institutions; governance indicators
JEL classification: E44; G24; H63
ΟΙ ΠΙΣΤΟΛΗΠΤΙΚEΣ ΑΞΙΟΛΟΓΗΣΕΙΣ ΚΑΙ ΤΑ ΘΕΜΕΛΙΩΔΗΜΕΓΕΘΗ ΤΗΣ ΕΛΛΗΝΙΚΗΣ ΟΙΚΟΝΟΜΙΑΣ
Δημήτρης ΜαλλιαρόπουλοςΔιεύθυνση Οικονομικής Ανάλυσης και Μελετώνκαι Πανεπιστήμιο Πειραιώς, Τμήμα Χρηματοοικονομικής και Τραπεζικής Διοικητικής
Πέτρος ΜηγιάκηςΔιεύθυνση Οικονομικής Ανάλυσης και Μελετών
ΠΕΡΙΛΗΨΗΤο παρόν άρθρο περιγράφει τις παραμέτρους των κρατικών πιστοληπτικών αξιολογήσεων καιαναπαράγει την ποσοτική συνιστώσα τους, εστιάζοντας στην περίπτωση της ελληνικής οικο-νομίας. Η πιστοληπτική διαβάθμιση της ελληνικής οικονομίας βρίσκεται χαμηλότερα από το όριοτης επενδυτικής κατηγορίας, όμως ορισμένα από τα θεμελιώδη μεγέθη της είναι καλύτερα απότα μέσα επίπεδα της κατηγορίας στην οποία ανήκει (ΒΒ) ή ακόμη και ανώτερων κατηγοριών.Με βάση την αναπαραγωγή του ποσοτικού σκέλους των κρατικών πιστοληπτικών διαβαθμίσεωντων τριών μεγάλων οίκων πιστοληπτικής αξιολόγησης, υποδεικνύεται ότι η βελτίωση της θέσηςτης ελληνικής οικονομίας στους δείκτες του θεσμικού περιβάλλοντος, σε επίπεδα παρόμοια εκεί-νων που είχαν καταγραφεί στις αρχές της δεκαετίας του 2000, θα συμβάλει σε ενδεχόμενη ανα-βάθμισή της στην επενδυτική κατηγορία.
51Economic Bulletin
July 2020 43
S OV ER E I GN C R ED I T R A T I NG S AND TH E F UNDAMENTA L S O F TH E G R E EK E CONOMY
51Economic BulletinJuly 202044
1 INTRODUCTION
Credit ratings are important inputs to portfolioallocation decisions, as they are widely used byinvestors as measures of default risk.2 Previousresearch has shown that both sovereign and cor-porate credit ratings are closely associated withthe level of risk premia in the underlying bonds.3
Understanding credit ratings and disentanglingtheir information is a crucial task both forinvestors and borrowers, thereby facilitating inan economically efficient way their decisions.
According to the methodologies of Credit Rat-ing Agencies (CRAs), credit ratings are assess-ments of the ability and the willingness of adebt issuer to pay back the debt in full. In theprocess of this assessment for sovereigns, CRAsanalyse several categories of fundamentals ofthe domestic economies, while the final ratingsalso incorporate judgment about the prospectsof the economy and potential developmentsthat pose upside or downside risks to the initialassessment.4 Sovereign credit ratings have awider importance for the national economy;they are linked to country risk assessments andthe so-called “country ceiling”, i.e. the maxi-mum rating that can be assigned to any entityof the public or the private sector originatingfrom the same economy.
In this paper, we estimate a model of sovereigncredit ratings based on the methodologies ofthe three major CRAs using data from 93countries over a long time span and then applyit to the data for Greece. This allows us todecompose Greece’s sovereign credit ratingsinto their main determinants and quantify thecontribution of each determinant to the over-all rating. Finally, we assess the importance ofinstitutional factors such as the quality of gov-
ernance for the prospects of an upgrade of theGreek sovereign rating to the investment gradecategory credit rating.5,6
The paper is organised as follows: The next sec-tion discusses the methodologies followed by thethree large CRAs (i.e. Fitch, Moody’s and Stan-dard and Poor’s) for rating sovereign entities.Section 3 compares the fundamentals of Greecewith those of sovereigns belonging to the sameand other rating categories. In Section 4 we pres-ent the econometric model and the estimationresults of credit scores, discuss the driving fac-tors of the Greek sovereign rating for the period2006-2018 and report the results of an impactanalysis around the prospect of an upgrade ofthe Greek sovereign rating to the investmentgrade category. Finally, Section 5 concludes.
2 SOVEREIGN CREDIT RATING METHODOLOGIES
The process of assigning credit ratings to sov-ereign entities followed by CRAs7 entails two
51Economic Bulletin
July 2020 45
S OV ER E I GN C R ED I T R A T I NG S AND TH E F UNDAMENTA L S O F TH E G R E EK E CONOMY 1
Dimitris MalliaropulosEconomic Analysis and Research Departmentand University of Piraeus, Department of Banking and Financial Management
Petros MigiakisEconomic Analysis and Research Department
1 Excellent research assistance by Ms Vanessa Kolonia is gratefullyacknowledged. The views expressed here are those of the authorsand not of the Bank of Greece or the Eurosystem. Any remainingerrors are ours.
2 For example, Morahan and Mulder (2013) find that four out of fiveinvestment managers use credit ratings in their portfolio allocationprocess.
3 For sovereign risk premia and ratings, see, among others,Livingston et al. (2010), Aizenmann et al. (2013), de Santis (2012),El-Shagi and von Schweinitz (2018) and Malliaropulos and Migiakis(2018); for corporate bonds, see Fons (1994), Longstaff et al.(2005), Heinke (2006) and Grothe (2013).
4 A number of studies criticise CRAs for providing inflated ratings,e.g. White (2010), Fulghieri et al. (2013) and Boermans and vander Kroft (2020). The issue of comparing credit ratings to “truemeasures of credit risk” is beyond the scope of this paper.
5 The investment grade category includes ratings equal to or betterthan BBB-/Baa3; ratings below this threshold (i.e. BB+/Ba1 orworse) are classified as sub-investment grade.
6 Since January 2020, Fitch’s sovereign credit rating for Greecestands at BB, Moody’s assigns a rating two notches lower than Fitch(i.e. B1) and Standard and Poor’s one notch lower (i.e. BB-).
7 In the present paper, we examine the three large, by internationalstandards, CRAs (in alphabetical order): Fitch, Moody’s andStandard and Poor’s (S&P).
stages: the quantitative or objective and thequalitative or subjective one. In the first (quan-titative) stage, each sovereign is assigned ascore or is ranked in relation to other sover-eigns, based on the country’s economic andpolitical fundamentals. In the second (quali-tative) stage, the quantitative score is adjustedusing experts’ opinions on the challenges oropportunities8 that the economy is expected toface in the near future. Usually, the qualitativeadjustment does not change the scoreassigned in the initial stage by more than oneto three notches.
During the first stage, CRAs incorporate intotheir quantitative analytical tools a distinct listof variables representing the fundamentals ofeach economy. For example, Fitch uses sixteenvariables as inputs to its sovereign ratingmodel. These are classified in four categories:structural features, macroeconomic perform-ance policies and prospects, public finances,and external finances (Fitch 2018). Similarly,Moody’s uses seventeen variables, which
51Economic BulletinJuly 202046
8 Political and geopolitical developments, nonlinearities in the publicdebt features and upcoming economic challenges are some of theissues considered at this stage.
Structural factors
• Governance
• GDP per capita
• GDP as % of world GDP
• Time since last default
• Broad money
Institutions and governance strength
• Quality of legislative and executive institutions
• Strength of civil society and the judiciary
• Fiscal policy effectiveness
• Monetary and macroeconomic policyeffectiveness
Political score
• Effectiveness, stability and predictability of policymaking, political institutions andcivil society
• Transparency and accountability of institutions, data and processes
• Debt payment culture
• External security risks
Macroeconomic factors
• Real GDP growth
• Volatility of real GDP
• Inflation rate
Economic strength
• Real GDP growth
• Volatility of real GDP
• Nominal GDP
• GDP per capita
Economic score
• GDP per capita
• Real GDP p/c trend growth
• Economic concentration and volatility
Public finances
• General government debt as % of GDP
• General government interest payments as % of revenues
• General government budget balance as % of GDP
• Foreign currency public debt as % of general government debt
Fiscal strength
• General government debt as % of GDP
• General government debt as % of revenues
• General government interest payments as % of revenues
• General government interest payments as % GDP
Fiscal score
• General government debt as % of GDP
• Change in net general government debt as % of GDP
• General government interest payments as % of revenues
• Contingent liabilities (financialinstitutions, public sector enterprises, off-budget contingent liabilities)
External finances
• Reserve currency status
• Sovereign net foreign assets
• Commodity dependence
• Reserves
• External interest service
• Current account balance plus FDI
Event risk
• Domestic political risk
• Ease of access to funding
• Risk of banking sector credit event
• Total domestic bank assets to GDP
• External vulnerability risk
External score
• Reserve currency status
• External liquidity (ratio of gross external financing to current account receipts (CAR) plusforeign exchange reserves)
• External indebtedness (ratio of net external debt to CAR)
Monetary score
• Exchange rate regime
• Monetary policy credibility
Fitch Moody’s Standard and Poor’s
Table 1 Variables used as input to the quantitative models of CRAs
belong to four categories: economic strength,institutions and governance strength, fiscalstrength, and susceptibility to event risk(Moody’s 2019). Finally, Standard and Poor’sincorporates sixteen factors in its model thatbelong to the following categories: politicalscore, economic score, external score, fiscalscore, and monetary score (Standard andPoor’s 2017).
While the terminology and the number of cat-egories imply that there is some deviationbetween the various quantitative models usedby CRAs, when we look closer into the indi-vidual variables that are included in each cat-egory, we find a remarkable similarity of thefactors taken into consideration for assigningsovereign credit ratings. Table 1 outlines theindividual variables used by each CRA in theirquantitative models.
The details of each category of variables usedin the process of the quantitative assessmentreveal the similarity of the factors assessedacross the three large CRAs. For example, realGDP growth, volatility of real GDP, GDP percapita, general government debt as a percent-age of GDP and general government interestpayments are used by all three CRAs. Institu-tional factors also play a prominent role in thequantitative assessment of all three CRAs.These factors reflect: (a) transparency andaccountability; (b) effectiveness of the admin-istration and political institutions; (c) the sov-ereign’s debt payment culture (Standard andPoor’s), the time since the last default (Fitch)or the government default history (Moody’s);(d) the quality of the legislation and the ruleof law; and (e) the perceived level of corrup-tion. In order to measure institutional factors,Fitch and Moody’s use the World Bank’sWorldwide Governance Indicators (WGIs);Fitch uses the average score of the six indi-vidual indicators, while Moody’s makes use ofthe indicators for regulatory quality and gov-ernment effectiveness.9
Overall, the factors taken into account in thestage of quantitative assessment of the credit
profile of each sovereign are very similar acrosscredit rating agencies. Also, the weightsassigned to the broad categories are similaracross the three CRAs: the most importantones are the institutional factors, followed byeconomic and fiscal factors, while external,monetary and event-risk factors are mostlyused for adjustment/calibration purposes.
Nevertheless, the three CRAS’ methodologi-cal frameworks bear some differences withrespect to the structure of their scorecards. Inparticular, the relative importance of the indi-vidual variables may differ due to differencesin their weighting schemes for individual vari-ables, while other variables, such as market-based indicators, are also taken into account byStandard and Poor’s and Moody’s. In practice,however, these differences do not result in sys-tematic divergences of more than two or threenotches in the final credit ratings assigned tothe same sovereign by the three CRAs.
3 HOW DOES THE GREEK ECONOMY COMPARETO OTHER ECONOMIES IN TERMS OF CREDITFUNDAMENTALS?
All CRAs rely on rankings of the values of thefundamentals vis-à-vis those of other sover-eigns of the same or other categories. Hence,it could be useful to compare the fundamentalvariables of the Greek economy with themedian of the rating categories in order toidentify both the strengths and the weaknessesof the Greek economy from the point of viewof rating agencies.
In order to do this, we analyse annual data ofthe individual variables used by CRAs for asample of 110 countries over the period 2006-2018.10 We only take into account categoriesand not notch-deviations within categories, e.g.
51Economic Bulletin
July 2020 47
9 However, in the qualitative stage Moody’s also makes use of threeadditional indicators: voice and accountability; rule of law; andcontrol of corruption.
10 The data set for 17 countries does not cover the entire period. So,while we use the data available for these countries in order toconstruct the rating buckets they belong to in each year, dependingon data availability, we exclude them from our econometricestimation (see Section 4).
sovereigns rated AA+ and AA- are included inthe AA category, sovereigns rated A+ and A-are included in the A category, and so on. Thecategories are dynamic, i.e. sovereigns migrateto higher or lower categories at the end of eachyear if they are upgraded or downgraded by atleast two CRAs. This ensures that the statisti-cal properties of the fundamentals in each cat-egory are representative of the rating categoryand not of specific groups of sovereigns.
In order to compare Greece’s fundamentalswith those of other sovereigns in the same orneighbouring rating categories, we compute themedian and the 10% and 90% quantiles of eachrating category. We then compare Greece’s fun-damentals with the median and the interquan-tile range of each rating category. In this way,we construct a statistical criterion similar to theone used by CRAs for classifying sovereignsinto rating categories before ranking their fun-damentals according to the weighting scheme ortaxonomy of their scorecard.
The way this statistical criterion works can beeasily understood: consider, for example, thecase where the value of a given fundamentalvariable for Greece, in a specific year, is bet-ter (worse)11 than the 90% (10%) quantile ofthe rating category where the country belongs.This would suggest that the country’s creditrating is likely to be upgraded (downgraded)based on this fundamental. The Appendixreports detailed charts of the median and theinterquantile range of a wide array of variablesused by CRAs for each rating class (AAA to B)based on the data of 110 countries in our sam-ple. The Appendix charts (A1-A15) also reportthe data for Greece and how they comparewith the data of other countries in each ratingclass. We briefly discuss the main observationsfrom these charts in the following paragraphs.
Institutional factors
One of the most important factors acrossCRAs is the quality of institutions and thepolitical landscape; this factor is measuredeither by individual or by aggregate indicators
of the quality of governance, such as the onesprovided by the World Bank’s Worldwide Gov-ernance Indicators (WGIs). Based on the aver-age of the six individual indicators provided bythe World Bank (see Chart A1 in the Appen-dix), Greece constantly ranks above themedian of its present rating category and evenbetter than the BBB category. However, thereis room for improvement, as Greece still liesbelow the median of sovereigns above single-A, which include other developed economiesand most euro area countries.
Chart 1 plots Greece’s ranking in each of the sixgovernance indicators reported by the WorldBank for the years 2001, 2008, 2012 and 2018.12
The chart shows that the position of Greece vis-à-vis the rest of the countries in the WorldBank’s governance indicators has deterioratedin the years following the global financial crisis.The greatest fall has been observed in the polit-ical stability indicator, which includes theabsence of violence, where Greece fell byalmost two deciles in the overall ranking, fromthe 55th percentile in 2008 to the 39th per-centile in 2012. Since then it has improved by11% in the percentile ranking, according to theWorld Bank’s 2018 WGI report.13 However,Greece remains lower than its ranking in 2001-2002,14 when it ranked at the 75th percentile, i.e.among the top 25% countries of the distribu-tion. Also, in the rule of law, Greece ranked atthe 59th percentile in 2018, which was evenlower than its ranking in 2012 (64th percentile),having fallen by about 15 percentiles since 2008.Similar, though smaller, falls have beenrecorded in the rest of the governance indica-tors, such as government efficiency, control ofcorruption, regulatory quality, and voice and
51Economic BulletinJuly 202048
11 We use the terms “better” and “worse”, instead of “larger” and“smaller”, as the direction of the effect of each variable on therating depends on the sign the variable has in the scorecard of theCRAs. For example, the debt-to-GDP ratio worsens the rating,whereas real GDP growth improves the rating.
12 Note that the rating agencies introduce the levels of the governanceindicators with a one-year lag, i.e. the ratings assigned in 2019 takeinto account the figures reported by the World Bank for 2018.
13 This report was released in September 2019 and is the most recentone.
14 Until 2002 the World Bank’s governance indicators were publishedevery two years. Greece’s historically highest ranking in the“political stability and absence of violence/terrorism” indicator wasrecorded in the 2001-2002 report.
accountability. Thus, a possible policy objectivefor Greece could be to regain the position thatthe country had in the World Bank’s governanceindicators before the eruption of the crisis.
Broad money is used by rating agencies tomeasure the level of financial intermediation in
the economies (Fitch 2018). As shown in ChartA2, Greece seems to be in an advantageousposition, relative to its current rating category(BB, including BB+ and BB-) on the basis ofthe broad money rating factor, as its level, whichalso bears a positive sign in the CRAs’ score-cards, is higher than that of the top 10% of the
51Economic Bulletin
July 2020 49
sovereign entities belonging to the same cate-gory. Moreover, on the basis of the same factor,Greece compares favourably even vis-à-vis theupper class of BBB-rated countries, with thevalue of the variable being close to the medianof the A category. This finding, combined withfact that Greece, as a member of the euro area,is assigned the highest score under the criterionof “reserve currency”, indicates the strongadvantages for its rating, stemming from itsmonetary and financial structure.
The institutional factors taken into account byCRAs also include GDP per capita as a meas-ure of the income level of an economy (seeChart A3). According to this criterion, Greecefares better than its current rating category (infact, above the BBB median). GDP as a per-centage of world GDP (as shown in Chart A4),another measure used by rating agencies to cap-ture the shock-absorption capacity of the econ-omy, provides a similar picture. However, therestructuring of public debt, which bears aheavy weight among the criteria used in CRAs’scorecards, poses a disadvantage for the Greekeconomy in the category of institutional factors.
Macroeconomic factors
The only advantage of the Greek economy inthe macroeconomic factors’ category is that oflower inflation, compared with the economiesincluded in almost all rating categories (seeChart A5); given the negative sign of this fac-tor, the low level of inflation in Greeceincreases its relative position. However, lowreal GDP growth (as shown in Chart A6) com-pared with other BB-rated sovereigns reducesthe probability of a rating upgrade. Finally,higher volatility of Greek GDP (see Chart A7)as a result of the deep recession of 2009-2013will continue to weigh on the probability of anupgrade, due to the fact that real GDP volatil-ity is measured over a 10-year period.
Fiscal factors
Fiscal consolidation has improved the pictureof Greek public finances. The positive primary
budget balance, achieved for the first time in2014, led Greece above the BB category eversince, despite the fact that the overall ratingof the country was much lower at that time.The continued effort has enabled Greece toexceed the median and interquantile range ofits rating category as well as that of sovereignsrated at the BBB category (as shown in ChartA8). Moreover, it brings Greece’s figureabove the medians of even upper ratingclasses, such as single-A and double-A ratingcategories.
By contrast, Greece’s high general governmentdebt-to-GDP ratio (as shown in Chart A9)exceeds the medians of all rating categories.Thus, as this factor carries a weight much heav-ier than that of the budget balance, it lowersthe total contribution of fiscal factors. This ispartially counterbalanced by the fact that theshare of public debt denominated in foreigncurrency is lower than both the BB and theBBB rating range, being close to the A-ratedmedian (see Chart A10). Finally, the high pub-lic debt ratio is also offset by reduced interestexpenses, due both to the lower coupon ratesof the more recent bond issues and to the lowcost of funding of the loans taken by Greecefrom the official sector (see Chart A11).
External factors
Last but not least, the external factors of theGreek economy reduce the prospects of ratingupgrades. In particular, the Greek State hasvery low net foreign assets (see Chart A12); theinterest service to foreign creditors is higherthan the BB-category median; the currentaccount balance, including FDI, as a ratio ofGDP is close to the BB median (see ChartA13); and the reliance of the Greek economyon one sector (tourism), reflected in the so-called “commodity dependence” factor, isstrong (see Chart A14). That said, it should benoted that the sovereign net foreign assets fac-tor does not account for the fact that part offoreign borrowing has been used by the GreekState to build a sizeable cash buffer (nearly10% of GDP at the end of 2019).
51Economic BulletinJuly 202050
4 THE GREEK SOVEREIGN CREDIT RATING
4.1 REPRODUCING THE qUANTITATIVE COMPONENTOF RATINGS
The aim of this section is to replicate the quan-titative component of ratings for Greece.15 Tothat end, we estimate an ordered probit modelfor 93 sovereigns worldwide using annual datafor the period from 2006 to 2018.16 In partic-ular, we estimated the following setup, equa-tion (1), which is based on the structure of
Fitch’s scorecard, as already presented inTable 1:
cit=a1+∑ki=1cs(xit)∙sit+∑k
i=1cm(xit)∙mit
+∑ki=1 cf (xit)∙fit+∑k
i=1cx (xit)∙xit+eit (1)
where cit is the credit rating assigned to sov-ereign i=1,2…N for each year t=1,2,…T, a1 is
51Economic Bulletin
July 2020 51
Global a1 Intercept5.042** (0.455)
None
Institutional iit
WB governance0.094** (0.003)
Rank
GDP per capita0.026** (0.002)
Rank
Share in world GDP0.643** (0.030)
Nat. log.
Default -1.667**
(0.226)Time since event
Broad money0.097
(0.084)Nat. log.
Macroeconomic mit
Real GDP volatility-0.556**
(0.069)Nat. log. (10y std.dev.)
Real GDP growth0.015
(0.015)3y ave.
Inflation -0.104**
(0.014)3y ave.
Fiscal fit
Gen.Gvt debt (%GDP)0.027** (0.002)
3y ave.
Gen.Gvt interest expenses(%rvn)
-0.019** (0.007)
3y ave.
Gen.Gvt budget balance(%GDP)
0.081**(0.009)
3y ave.
Foreign currency public debt -0.015**
(0.002)3y ave.
External xit
Reserve currency status0.587** (0.042)
Eval.
Sovereign net foreign assets0.011** (0.001)
%CXR
Commodity dependence-0.003
(0.003)3y ave.
Reserves0.047** (0.011)
%CXP
External interest service-0.004
(0.001)3y ave.
Category K={x1,x2,…xi} Variable xit Estimated coefficient C(xit) Transformation F(xit)
Table 2 Ordered probit estimation results
Notes: The table presents the estimated coefficients of the individual variables of the ordered probit model described in equation (1), with thecredit ratings of 93 sovereigns as the dependent variable. The sample is 2006-2018. The final column to the right describes the way the variableis transformed in order to be incorporated into the reproduced scorecard. CXR: current account receipts. CXP: current external payments. Asterisks (** and *) denote significance (at the 1% and 5% level, respectively).
15 Previous studies aimed at quantifying the effect of factors on sov-ereign ratings include inter alia Afonso et al. (2009), D’Agostinoand Lennkh (2016), Brůha et al. (2017) and Lennkh and Mosham-mer (2018).
16 The source of the data is Fitch Ratings.
the global intercept, sit is the vector of the insti-tutional/structural variables, mit is the vector ofmacroeconomic variables, fit is the vector of fis-cal variables and xit is the vector of externalvariables, with k being the number of variablesin each category.17 Finally, eit is the panel dataresidual from the estimation. Table 2 presentsthe estimated coefficients for each variable andthe transformation of each variable, asdescribed by Fitch in its sovereign ratingsmethodology.
Based on the estimation of the coefficients, asshown in Table 2, we can reproduce the score-card of Fitch by taking into account the coef-ficient c(xit ) and the transformation F(xit ) ofeach variable. In particular, the score of eachvariable is calculated as follows:
Score(xit )=c(xit )∙F(xit ) (2)
Similarly, the score for each category of vari-ables is the sum of its individual variables:
Score(Kt )=∑ki=1c(xit )∙F(xit ) (3)
Finally, the aggregate score of each country isthe sum of the scores of the categories and theglobal intercept:
Score=α+∑4j=1 Score(Kj) (4)
We use the aggregate score to rank each coun-try and produce its initial rating. To do so, weuse a rule that associates each score with a rat-ing category, as shown in Table 3.
The replication of the scores of the fundamen-tals of the Greek economy, based on the abovesetup, facilitates both the monitoring of devel-opments in the sovereign credit rating of Greeceand the quantification of the impact of past andexpected or assumed developments in Greece’sfundamentals. Also, it enables the estimation ofthe contribution of each individual variable tothe quantitative component of the final rating.
Chart 2 plots Greece’s estimated score alongwith the actual rating (computed as the aver-
age rating assigned by the three CRAs; for theratings of individual CRAs, see Chart A15 inthe Appendix) as well as the individual contri-bution of each variable over the 2006-2018period. The chart shows that the estimatedscore for Greece follows closely the averagesovereign credit rating assigned to the countryby the three large CRAs. The two lines followeach other in close connection until 2009,18
indicating that the rating assigned to Greeceuntil the outbreak of the crisis was largely inline with the economy’s fundamentals, as meas-ured by the quantitative component (score) of
51Economic BulletinJuly 202052
ΑΑΑ>15.5 ΑΑΑ>15.5
15.5>ΑΑ>12.5
15.5>ΑΑ+>14.5
14.5>ΑΑ>14
13.5>ΑΑ+>12.5
12.5>Α>9.5
12.5>Α+>11.5
11.5>Α>11
10.5>Α->9.5
9.5>ΒΒΒ>6.5
9.5>ΒΒΒ+>8.5
8.5>ΒΒΒ>8
7.5>ΒΒΒ->6.5
6.5>ΒΒ>3.5
6.5>ΒΒ+>5.5
5.5>ΒΒ>5
4.5>ΒΒ->3.5
3.5>Β>0.5
3.5>Β+>2.5
2.5>Β>2
1.5>Β->0.5
0.5>C/D 0.5>C or D
Table 3 Translating scores into ratings
17 The category of institutional factors includes the followingvariables: the average of the World Bank’s WorldwideGovernance Indicators, GDP per capita, GDP as % of world GDP,broad money, and time since last default. The economicactivity/macroeconomic factor includes: real GDP growth, realGDP volatility and the annual rate of change in CPI. The fiscalfactor includes: general government debt as % of GDP, generalgovernment interest payments as % of revenues, generalgovernment budget balance as % of GDP and foreign currencypublic debt as % of general government debt. Finally, the externalfactor includes: the status of reserve currency, sovereign net foreignassets as % of GDP, the degree of commodity dependence, the levelof foreign exchange reserves, the external interest service and thecurrent account balance plus FDI.
18 Greece was rated at A+ on average by the three CRAs during 2009;by the end of that year a downgrade cycle had begun, whichescalated with the restructuring of the Greek public debt in 2012.Ever since, with a short interruption in 2015, CRAs havecontinuously upgraded Greece’s sovereign credit rating.
CRAs’ ratings.19 After 2010, the actual ratingof Greece drops two to three notches below theestimated score, which measures the effect offundamentals on the rating. This divergence isdue to the judgmental component, whichlargely captures the effect of the Greek debtrestructuring in March 2012.20 An alternativeinterpretation of the divergence between actualratings and estimated scores after 2010 is thatCRAs overreacted to the deterioration of eco-nomic fundamentals after the sovereign debtcrisis erupted. Distinguishing between thesetwo explanations is difficult and certainlybeyond the scope of this paper. Nevertheless,the fact that this divergence persists after 2010suggests that it is driven by the debt restruc-turing rather than by an overreaction of CRAsto the deterioration of economic fundamentalsafter the eruption of the sovereign debt crisis.
4.2 THE DRIVERS OF SOVEREIGN RATING CHANGESFOR GREECE, 2006-2018
The Greek sovereign credit rating stoodfirmly within the investment grade categoryfrom the late 1990s, i.e. before the country’saccession to the European Monetary Union,until the global financial crisis, when a down-grade cycle began for several euro area mem-bers.21 This downward revision of the creditprofiles of euro area countries impaired theconditions of refinancing their debt in the
51Economic Bulletin
July 2020 53
19 Our findings are in line with those reported in Lennkh andMoshammer (2018) for Greece.
20 In March 2012, Greece restructured EUR 205 billion of public debt(165% of Greek GDP). Private investors suffered a 53% haircuton the face value of their Greek bond holdings.
21 Brůha et al. (2017) attribute this wave of downgrades following theglobal financial crisis to a structural break that led to greaterimportance of the quantitative stage and less optimism in thejudgmental stage.
bond market and marked the beginning of theeuro area debt crisis.
The downgrades of the Greek sovereign creditrating had a prominent role in this regard.This is because the Greek State was the firstamong euro area countries to lose its invest-ment grade status. As shown by El-Shagi andvon Schweinitz (2018), such downgrades to thesub-investment grade status can lead to a per-sistent increase in the cost of funding thatcould jeopardise public debt sustainability.This rise in the cost of funding, as well as theresulting reduction of financial flows (e.g.those reflected in the portfolio holdings of theinternational investment position) may resultin an accentuation of the downturn of the eco-nomic cycle, thus creating adverse feedbackloops between credit ratings and economicfundamentals.22
In the case of Greece, as shown in Chart 2, theinitial downgrades during and immediately afterthe global financial crisis were largely related tothe deterioration of the economy’s fundamen-tals. In particular, the deterioration of the esti-mated score between end-2008 and end-2010largely explains the loss of the investment gradestatus for Greece; over this period, the esti-mated score declines by 3.5 points, which isequivalent to a downgrade of four notches. Still,this figure, though informative, does not pro-vide an answer to the question which one of themany adverse developments in the period from2008 and up to 2010 had a greater contributionto the loss of the investment grade status.
In order to get a better understanding of theunderlying drivers of ratings downgrades and
51Economic BulletinJuly 202054
22 See among others Gibson et al. (2017) and Amato and Furfine (2004).
upgrades, we calculate the contributions of theunderlying variables across specific sub-periodsof the sample. We separate the sample in threesub-periods: from 2009 to 2011, 2012, and from2013 to 2018. Separating the sample in threesub-samples rather than two (e.g. 2009 to 2012and 2012 to 2018) allows to isolate the effectof the debt restructuring of 2012 on the ratingfrom the effect of fundamentals. Chart 3 illus-trates the contributions to Greece’s sovereignrating score of each group of fundamentals.
The chart shows that the factors contributingto the downgrades over the 2009-2011 periodare different from the ones driving theupgrades during the 2013-2018 period. Morespecifically, the external and fiscal imbalancescombined have contributed to a reduction inGreece’s sovereign rating score by around -2.4
(which is equivalent to a downgrade of threenotches) vis-à-vis -1.5 score units, due to thedeterioration of macroeconomic (mainly) andinstitutional (secondarily) factors. The devel-opment of the scores in the institutional factorscategory in 2012 mainly reflects the debtrestructuring, as this category includes the vari-able “time since default”. Finally, theupgrades observed over the 2013-2018 periodare primarily attributed to the improvement ofthe institutional factors, which have added 1.82score unit and, secondarily, to the improve-ment in the macroeconomic environment(+1.02 score unit), with the external and fiscalfactors adding 0.52 and 0.42 unit, respectively.
The improvement of the score of institutionalfactors after 2012 reflects to some extent theincreasing distance from the time of debt
51Economic Bulletin
July 2020 55
restructuring. More importantly though, itreflects improved governance, as suggested bythe rise of Greece’s ranking on the “politicalstability and absence of violence” sub-index ofthe World Bank’s governance indicators. Nev-ertheless, there is room for further improve-ment, if we compare Greece’s ranking in 2018on all individual governance indicators with itsposition before the crisis.
4.3 HOW CAN GREECE BE UPGRADED TO THEINVESTMENT GRADE?
In the present section, we report the results ofa simulation analysis of potential develop-ments, with the aim to inform how the Greeksovereign credit rating can be upgraded to theinvestment grade.23 The simulations assume (a) that Greece’s institutional score improves,which positions the country on the basis of theWorld Bank’s governance indicators to its his-torical high observed in the early 2000s, and (b) that one more year passes without anycredit event. All other variables remain at thelevels assumed by the CRAs in their mostrecent updates.24
Chart 4 presents the results of this simulationexercise. The results indicate that, all elseequal, the improvement in governance to 2001levels along with one more year passing with-out a credit event would improve Greece’sscore enough to contribute to an upgrade of itsrating to investment grade (i.e. equal to orabove BBB-), provided that the adjustment atthe second stage remains at the present level.This highlights the importance of the institu-tional environment for the sovereign credit rat-ing of the Greek economy and, as a conse-quence, for entities of both the public and theprivate sector, whose cost and opportunities offunding are associated with country risk.
5 CONCLUDING REMARKS
We estimate the quantitative component ofsovereign credit ratings following methodolo-gies of the three major Credit Rating Agen-cies. We then use the model to replicateGreece’s sovereign credit ratings over the2006-2018 period. We show that Greek sov-ereign ratings over this period have largely fol-lowed the economy’s fundamentals. However,we find that, after 2010, Greece’s actual ratingdrops two to three notches below the esti-mated score, which measures the effect of fun-damentals on the rating. This divergence is dueto the judgmental component of ratings,which, according to our interpretation, largelycaptures the effect of the Greek debt restruc-turing in March 2012.
At present, the Greek economy has severaladvantages and few disadvantages, comparedwith economies belonging to the same cate-gory (BB). On the positive side, Greece out-performs its BB peers due to the strength ofits institutions, the developed status of itseconomy, the strength of the monetary regimeand its high income per capita. On the nega-tive side, Greece lags behind its past per-formance in terms of quality of governance,which constitutes one of the most importantrating factors for sovereign entities. In fact,our simulations suggest that improving thequality of governance to pre-crisis levels is anecessary condition for an upgrade of theGreek sovereign credit rating to investmentgrade.
51Economic BulletinJuly 202056
23 In order to regain the investment grade status, Greece has to beupgraded by two notches by Fitch, three notches by Standard andPoor’s and four notches by Moody’s.
24 See the rating action reports by Fitch (24 July 2020), Moody’s (10July 2020) and S&P (24 April 2020).
Afonso, A., P. Gomes and P. Rother, (2009), “Ordered response models for sovereign debt rat-ings”, Applied Economics Letters, 16(8), 769-773.
Aizenmann, J., M. Binici and M. Hutchinson (2013), “Credit ratings and the pricing of sover-eign debt during the euro crisis”, Oxford Review of Economic Policy, 29, 582-609.
Amato, J.D. and C.H. Furfine (2004), “Are credit ratings procyclical?”, Journal of Banking andFinance, 28(11), 2641-2677.
Boermans, M.A. and B. van der Kroft (2020), “Inflated credit ratings, regulatory arbitrage andcapital requirements: Do investors strategically allocate bond portfolios?”, De NederlandscheBank Working Paper No. 673.
Brůha, J., M. Karber, B. Pierluigi and R. Setzer (2017), “Understanding sovereign rating move-ments in euro area countries”, European Central Bank Working Paper No. 2011.
D’ Agostino, A. and R.A. Lennkh (2016), “Euro area sovereign ratings: an analysis of fundamentalcriteria and subjective judgment”, European Stability Mechanism Working Paper No. 14.
De Santis, R. (2012), “The euro area sovereign debt crisis: safe haven, credit rating agencies andthe spread of the fever from Greece, Ireland and Portugal”, European Central Bank WorkingPaper No. 1419.
El-Shagi, M. and G. von Schweinitz (2018), “The joint dynamics of sovereign rating and gov-ernment bond yields”, Journal of Banking and Finances, 97, 198-218.
Fitch (2018), Sovereign rating criteria (publication date: 19 July 2018). Fons, J.S. (1994), “Using default rates to model the term structure of credit risk”, Financial Ana-lysts Journal, 50(5), 25-33.
Fulghieri, P., G. Strobl and H. Xia (2013), “The economics of solicited and unsolicited creditratings”, Review of Financial Studies, 27(2), 484-518.
Gibson, H., S. Hall and G. Tavlas (2017), “Self-fulfilling dynamics: the interactions of sovereignspreads, sovereign ratings and bank ratings during the euro financial crisis”, Journal of Inter-national Money and Finance, 73B, 371-385.
Grothe, M. (2013), “Market pricing of credit rating signals”, European Central Bank WorkingPaper No. 1623.
Heinke, V.G. (2006), “Credit spread volatility, bond ratings and the risk reduction effect of watch-listings”, International Journal of Finance and Economics, 11(4), 293-303.
Lennkh, R.A. and E. Moshammer (2018), “An analysis of the degree, changes and source ofMoody’s judgment”, European Stability Mechanism Working Paper No. 27.
Livingston, M., J. Wei and L. Zhou (2010), “Moody’s and S&P ratings: are they equivalent? Con-servative ratings and split rated bond yields”, Journal of Money, Credit and Banking, 42(7), 1267-1293.
Longstaff, F.A., M. Sanjay and E. Neis E. (2005), “Corporate yield spreads: default risk or liq-uidity? New evidence from the credit default swap market”, Journal of Finance, 60(5), 2213-2253.
Malliaropulos, D. and P. Migiakis (2018), “The re-pricing of sovereign risks following the GlobalFinancial Crisis”, Journal of Empirical Finance, 49, 39-56.
Moody’s (2019), Sovereign ratings methodology (publication date: 25 November 2019).Morahan, A. and C. Mulder (2013), “Survey of reserve managers: lessons from the crisis”, Inter-
national Monetary Fund Working Paper No. 13/99.Standard and Poor’s (2017), Sovereign rating methodology (publication date: 18 December 2017).White, L.J. (2010), “Markets: The credit rating agencies”, Journal of Economic Perspectives,
24(2), 211-226.
51Economic Bulletin
July 2020 57
R E F E R ENC E S
51Economic BulletinJuly 202058
APP END I X
51Economic Bulletin
July 2020 59
51Economic BulletinJuly 202060
51Economic Bulletin
July 2020 61
51Economic BulletinJuly 202062
51Economic Bulletin
July 2020 63
51Economic BulletinJuly 202064
51Economic Bulletin
July 2020 65
51Economic BulletinJuly 202066
51Economic Bulletin
July 2020 67
51Economic BulletinJuly 202068
51Economic Bulletin
July 2020 69
51Economic BulletinJuly 202070
51Economic Bulletin
July 2020 71
51Economic BulletinJuly 202072
Constantina BackinezosEconomic Analysis and Research Department
Stelios PanagiotouEconomic Analysis and Research Department
Christos PapazoglouEconomic Analysis and Research Department and Panteion University, Department of International, European and Area Studies
ABSTRACTThe substantial narrowing of the Greek current account deficit ―by more than 13 percentagepoints of GDP― constitutes a significant part of the overall adjustment of the economy in the2008-2019 period. A major issue, however, is related to whether the factors driving this exter-nal correction are temporary, due mostly to the economic cycle, or structural, in which case theyare expected to be sustained over the medium term. In this paper we evaluate the degree to whichcyclical versus structural developments have driven the pattern of correction of the country’s exter-nal imbalance, as this is essential in assessing its evolution going forward. Our analysis shows thatthe current account adjustment has in a considerable part been driven by structural factors, whichaccording to our baseline scenario account for 60% of the overall adjustment. Therefore, mostof the correction is of a permanent nature, which means that the relatively high deficits of thepre-crisis period are not expected to re-emerge any time soon. Our results are robust to variousincome elasticities of trade; the non-cyclical component ranges from 47% to 74% of the currentaccount deficit correction.
Keywords: cyclically adjusted current account; macroeconomic imbalances; output gap; currentaccount; business fluctuations
JEL classification: F32; F40; E32
51Economic Bulletin
July 2020 73
THE CURR ENT A C COUNT AD J U S TMENT I N G R E E C EDUR I NG TH E C R I S I S : C Y C L I C A L O R S T RUC TURA L ?
Κωνσταντίνα ΜπακινέζουΔιεύθυνση Οικονομικής Ανάλυσης και Μελετών
Στέλιος ΠαναγιώτουΔιεύθυνση Οικονομικής Ανάλυσης και Μελετών
Χρήστος ΠαπάζογλουΔιεύθυνση Οικονομικής Ανάλυσης και Μελετώνκαι Πάντειο Πανεπιστήμιο, Τμήμα Διεθνών, Ευρωπαϊκών και Περιφερειακών Σπουδών
ΠΕΡΙΛΗΨΗΗ σημαντική μείωση του ελλείμματος του ισοζυγίου τρεχουσών συναλλαγών της Ελλάδος ―πάνωαπό 13 ποσοστιαίες μονάδες του ΑΕΠ― αποτελεί κεντρικό παράγοντα της συνολικής προ-σαρμογής της οικονομίας την περίοδο 2008-2019. Ένα μείζον ερώτημα, ωστόσο, σχετίζεται μετο αν οι παράγοντες που οδήγησαν σ’ αυτή τη διόρθωση της εξωτερικής ανισορροπίας είναι προ-σωρινοί, αποτέλεσμα κυρίως του οικονομικού κύκλου, ή διαρθρωτικοί, οπότε αναμένεται να δια-τηρηθούν μεσοπρόθεσμα. Στο παρόν άρθρο εξετάζουμε τη σχετική σημασία των κυκλικών καιτων διαρθρωτικών παραγόντων που συνέβαλαν στη διόρθωση της ανισορροπίας του εξωτερι-κού τομέα της χώρας. Αυτό είναι απαραίτητο για την αξιολόγηση της εξέλιξης του εξωτερικούτομέα μεσοπρόθεσμα. Η ανάλυσή μας δείχνει ότι η προσαρμογή του ισοζυγίου τρεχουσών συναλ-λαγών οφείλεται σε μεγάλο βαθμό σε διαρθρωτικούς παράγοντες, οι οποίοι σύμφωνα με τοβασικό σενάριό μας αντιστοιχούν στο 60% της συνολικής προσαρμογής. Επομένως, το μεγα-λύτερο μέρος της διόρθωσης είναι μόνιμου χαρακτήρα, πράγμα που σημαίνει ότι τα σχετικάυψηλά ελλείμματα της περιόδου πριν από την κρίση δεν αναμένεται να επανεμφανιστούνσύντομα. Τα αποτελέσματά μας εξακολουθούν να ισχύουν σε διάφορες τιμές εισοδηματικής ελα-στικότητας των εμπορικών συναλλαγών και η μη κυκλική συνιστώσα της προσαρμογής κυμαί-νεται από 47% έως 74% της συνολικής διόρθωσης του ελλείμματος του ισοζυγίου τρεχουσώνσυναλλαγών.
51Economic BulletinJuly 202074
Η ΠΡΟΣΑΡΜΟΓΗ ΤΟΥ ΕΛΛΕΙΜΜΑΤΟΣ ΤΟΥ ΙΣΟΖΥΓΙΟΥΤΡΕΧΟΥΣΩΝ ΣΥΝΑΛΛΑΓΩΝ ΤΗΣ ΕΛΛΑΔΟΣ ΚΑΤΑ ΤΗ ΔΙΑΡΚΕΙΑ ΤΗΣ ΚΡΙΣΗΣ: ΚΥΚΛΙΚΗ Ή ΔΙΑΡΘΡΩΤΙΚΗ;
1 INTRODUCTION
The adjustment of the Greek economy inrecent years is primarily linked with the sig-nificant correction of large macroeconomicimbalances generated in the run-up to the eco-nomic crisis of 2008 and includes, inter alia, amajor improvement in the external sector. Inparticular, the reduction of the current accountdeficit during the crisis was indeed remarkable,since it fell from 15.1% of GDP in 2008 to1.4% of GDP by the end of 2019 (see Chart 1).This undoubtedly constitutes a success storyfor the Greek economy.
An important issue ―which has drawn a lot ofattention recently2― refers to the extent towhich this adjustment will be sustained overthe medium term. This depends on the mag-nitude of the part of the external correctionthat can be attributed to the economic cycle,relative to the part that was the result of struc-tural factors, which are independent of thecycle and, therefore, more permanent innature. In particular, the cyclical part is pri-marily related to the large negative output gapof the post-crisis period. This in turn was theresult of the large decline in domestic demand,which, through a considerable drop inimports, contributed to the reduction of thecurrent account deficit. Moreover, the cor-rection associated with cyclical factors is theresult of the deeper recession of the Greekeconomy in comparison with the economies ofits major trading partners, a differential thatallowed imports to fall faster than exports, low-ering the Greek current account deficit. Thesize of such cyclical adjustment is what raises
concerns about the sustainability of the over-all improvement in the recovery phase of thecycle. That is, as the economy recovers and thecountry’s real GDP approaches its potentiallevel, raising the prospect of a positive outputgap in the medium term, the likelihoodincreases that the large external deficits of thepast might re-emerge.
On the other hand, to the extent that a signif-icant part of the current account correction canbe attributed to structural factors, theachieved adjustment will be permanent. Thatis, structural factors are, among other things,associated with competitiveness gains as aresult of productivity growth, improved fiscalpositions, stronger institutions, a shift of pro-duction towards tradable sectors, financialdeepening and favourable demographicchanges. The evolution of these factors couldbe of significant importance in capturing broadmedium-term trends in external imbalances.For instance, to the extent that the decline inthe current account deficit is attributable to ris-ing exports, resulting from competitivenessgains, it reflects structural changes and isexpected to persist over the medium term.
Consequently, the answer to the question onwhether the current account deficit will even-tually return to its pre-crisis level is criticallylinked with the nature (cyclical or structural)of the achieved external sector correction. Theaim of this paper is to investigate which part of
51Economic Bulletin
July 2020 75
THE CURR ENT A C COUNT AD J U S TMENT I N G R E E C EDUR ING THE CR I S I S : C YC L I C A L OR S TRUCTURAL ? 1
Constantina BackinezosEconomic Analysis and Research Department
Stelios PanagiotouEconomic Analysis and Research Department
Christos PapazoglouEconomic Analysis and Research Department and Panteion University, Department of International, European and Area Studies
1 The authors would like to thank H. Balfoussia for her usefulcomments and invaluable insights. The views expressed are of theauthors and do not necessarily reflect those of the Bank of Greece.The authors are responsible for any errors or omissions.
2 For example, see IMF (2019).
the accomplished adjustment is the result ofcyclical factors and thus will fade away in themedium term and which part is due to struc-tural factors and therefore is more likely tostay. In Section 2 we examine in more detailrecent developments in the Greek currentaccount. The methodology and data employedare explained in Section 3. The results of theanalysis and their robustness check are dis-cussed in Sections 4 and 5, respectively. Finally,the concluding remarks of the analysis are pre-sented in Section 6.
2 DEVELOPMENTS IN THE GREEK CURRENTACCOUNT
As pointed out, the current account deficitreached its peak of EUR 36.6 billion (15.1% ofGDP) in 2008 at the outbreak of the globalfinancial crisis and just before the collapse ofworld trade that followed in 2009 (see Chart 1).The lead-up to the economic crisis (i.e. 2002-2008) was characterised by low private savings,which could not meet the investment needs ofthe economy and, combined with extensive fis-cal imbalances, led to the build-up of sizeableexternal imbalances.3 In terms of the trade bal-ance, there had been a rising goods deficit, onlypartially counterbalanced by the services(mainly travel and sea transport) surplus,which remained rather stable at a relatively lowlevel. As an overall assessment, the country’sentry into the Monetary Union increaseddomestic demand which, reinforced further bya pro-cyclical fiscal policy, pushed upwardsimports and the trade deficit, while productiveresources were to a great extent trapped intonon-tradable sectors.
At the same time, exports of goods were risingand gaining market share, as Greek exportersmanaged to exploit their position in the South-east European markets,4 compensating for theloss in cost/price competitiveness (largely dueto competition from Asian countries) and forthe less favourable product composition ofGreek exports (which mainly consisted of low-to medium-technology products).5 Despite
that, exports remained at approximately one-third of imports and their expansion was notfast enough to outweigh the increasing goodsbalance deficit. It should also be noted thatexports and investment (concentrated mostlyin housing construction) have a considerableimport content of approximately 31% and41%, respectively.6
The decade after the crisis (i.e. 2009-2019) hasbeen characterised by a substantial improve-ment in the Greek external balance. The cur-rent account deficit dropped by 13.7 percent-age points (pp) and reached 1.4% of GDP in2019. Most of the adjustment can be attributedto exports of goods, which increased to 17.3%of GDP from 9.1% of GDP in 2008. Shrinkingdomestic demand urged domestic producers tointensify their efforts to enter and expand intoforeign markets. At the same time, exports ofservices adjusted upwards, albeit at a weakerpace due to the underperformance of seatransport and the slow recovery of tourism.Overall, since the end of 2009, exports of goodsand services excluding sea transport haveincreased by 64% at constant prices (whileexports of goods increased by almost 62% overthe same period). At the same time, importshave dropped by 30% since their peak in 2008.Imports started rising again after 2016 as aresponse to the improvement of economicactivity and the need for replacing capitalequipment, which raises concerns overwhether the current account improvement canbe sustained if growth in all components ofdomestic demand and exports accelerates.However, despite the fact that imports areneeded for growth, there have been some sub-stantial structural changes in the Greek econ-omy during this adjustment, especially on theside of exports, that could help sustain theexternal rebalancing achieved.
The recovery in competitiveness is complete interms of cost and almost complete in terms of
51Economic BulletinJuly 202076
3 See Bank of Greece (2020). 4 See Papazoglou (2007).5 See Backinezos et al. (2019) and Athanasoglou et al. (2010).6 See Bank of Greece (2018a).
price, while some substantial improvement innon-price competitiveness has also beenachieved. Greek exports, particularly non-oilexports, lost their market share in the begin-ning of the crisis but started regaining andmaintaining their position in internationalmarkets since 2016. Continuing the imple-mentation of structural reforms in productmarkets and institutions will further increasecompetitiveness, especially non-price. Inaddition, while the sectoral distribution ofGreek exports has not changed considerably,it has started contributing in a positive way tothe expansion into foreign markets, allowingexporters to benefit from the prolonged rise incost/price competitiveness.7
Domestic firms have become more export-ori-ented, although they went through a period ofeconomic contraction.8 In many sectors, the
proportion of domestic output exported hasincreased and it is most unlikely that theseexporters will retract from their positions ininternational markets in favour of the domes-tic market. Furthermore, a considerable shareof Greek exports relies on products that areless elastic to foreign demand changes (such asfood and pharmaceuticals which represent26% and 9% of total non-oil exports, respec-tively) and therefore more resilient to foreigndemand changes, as the reaction to the recent(2019) slowdown in the world economy hasproven.
With regard to tourism, its performance in2019 showed that the image of the country hasimproved significantly, since a considerable
51Economic Bulletin
July 2020 77
7 See Backinezos et al. (2019).8 See Bank of Greece (2017).
rise in average expenditure per trip wasrecorded, reflecting a shift towards higher-income travellers, in contrast with develop-ments in past years when travel receipts weredriven mainly by the large numbers of arrivals.9
Besides, since 2002 there has been a qualityimprovement in the Greek-controlled fleet,which since 2011 is younger than or on a parwith the world fleet. This fleet renewal wassupported by the exceptionally advantageousfreight markets in 2007-2008 and the avail-ability of ship finance in that period. Moreover,Greek-controlled shipping expanded into newsectors such as the carriage of Liquefied Nat-ural Gas that required fresh investment intechnologically advanced vessels (Greek Ship-ping Co-operation Committee). Finally, on theback of all those changes, exports grew as aproportion of imports leading to the contrac-tion of the goods and services balance deficit,which accounts for the bulk of the currentaccount. Also, such changes are more struc-tural in nature and are expected to have morepermanent effects on the export performanceof the Greek economy.
3 METHODOLOGY AND DATA
In the literature, there are two main streams inanalysing the adjustment of the currentaccount balance, in order to isolate the cycli-cally adjusted (or structural) current account.The first stream is based on the intertemporalapproach of the current account and, relyingon a variety of parameters such as macroeco-nomic variables, demographics, the financial,political and institutional environment, etc.,directly estimates the structural component ofthe current account. It is mostly employed byinternational institutions through the use ofcountry panel data.10 The second stream cal-culates the cyclically adjusted trade balanceindirectly after estimating the cyclical compo-nent. The estimation of the latter is based onthe output gaps of home and trading partners,as well as on the income elasticities of importsand exports.11 In our analysis, we follow themethodology employed in the second stream.
Specifically, our methodology closely followsHaltmaier (2014) and is augmented with Fabi-ani et al. (2016). According to the IMF (2009),“the current account shows flows of goods,services, primary income, and secondaryincome between residents and nonresidents”.Therefore, in nominal terms, it is presented as:
(1)
where:
ca: current account balance (nominal);
x: exports of goods and services (nominal);
m: imports of goods and services (nominal);
bpi: balance of primary income (e.g. compen-sation of employees, dividends, interest, etc.,nominal); and
bsi: balance of secondary income (e.g. personaltransfers, current international assistance, etc.,nominal).
Our objective is to separate the cyclical fromthe non-cyclical element of the currentaccount. In our analysis, the adjustment con-cerns exclusively the exports and imports ofgoods and services, i.e. the trade balance.
We first define:
ΔΧ (ΔΜ) as the difference between the poten-tial and the observed level of real exports(imports), i.e. the cyclical component of theexports (imports).
Then from the export side of the trade balancewe have:
(2)
where:
51Economic BulletinJuly 202078
, -. # ./ % . or ./ # . ' -. (2)
0 , (3) 2 (4)
. , (5)
9 , 5 , (6)
2 2 , (7)
. , (8) - , , , (9)
- , , , , , , , (10)
$ , (11)
- , , , , , , (12)
& , (13)
, , , ,
, , ,
, , , (14)1
, , , (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
!" # $ % &' ()* ' (+*,, - . (2)
0 , (3) 2 (4)
. , (5)
9 , 5 , (6)
2 2 , (7)
. , (8) - , , , (9)
- , , , , , , , (10)
$ , (11)
- , , , , , , (12)
& , (13)
, , , ,
, , ,
, , , (14)1
, , , (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
9 See Bank of Greece (2019a). 10 See for instance the IMF’s External Balance Assessment (EBA)
methodology in Phillips et al. (2013).11 See for instance Haltmaier (2014).
X*: potential exports (real), i.e. the level ofexports consistent with trading partners’ poten-tial output; and
X: observed exports (real).
Let’s define as θX (θM) the elasticity of realexports (imports) with respect to the real GDPof trading partners (of the home economy).
(3)
where:
Yf: trading partners’ real GDP; and
ΔΥf: the difference between the potential realGDP (Y*f) and the observed real GDP of trad-ing partners.
Equation (3) can be rewritten as:
(4)
Substituting equation (4) in equation (2), we get:
(5)
The output gap is defined as:
and it can be rearranged as:
(6)
Rearranging the last element of equation (5)
and substituting equation (6), we get:
(7)
Finally, we substitute (7) into (5) and we get:
(8)
or
(9)
If we assume that the export prices (PX) remainunchanged, the cyclical component of exportsin nominal terms is:
(10)
and the corresponding cyclically adjusted nom-inal exports are:
(11)
Accordingly, the equations for nominalimports are as follows:
Cyclical component of imports in nominal terms:
(12)
Cyclically adjusted nominal imports:
(13)
Therefore, the cyclically adjusted nominaltrade balance is given by:
(14)12
It follows from equation (1) that the cyclicallyadjusted current account (in nominal terms)will be:
(15)
Finally, the current account as a percentage ofnominal GDP (Ynom) is:
(16)
51Economic Bulletin
July 2020 79
, - (2)
01 # 2343256456, (3) 2 (4)
. , (5)
9 , 5 , (6)
2 2 , (7)
. , (8) - , , , (9)
- , , , , , , , (10)
$ , (11)
- , , , , , , (12)
& , (13)
, , , ,
, , ,
, , , (14)1
, , , (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
, - (2)
0 , (3) 231 = 01 25
656 (4)
. , (5)
9 , 5 , (6)
2 2 , (7)
. , (8) - , , , (9)
- , , , , , , , (10)
$ , (11)
- , , , , , , (12)
& , (13)
, , , ,
, , ,
, , , (14)1
, , , (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
, - (2)
0 , (3) 2 (4)
./ # . ' -. # 7, 8 ' 01 256
56 (5)
9 , 5 , (6)
2 2 , (7)
. , (8) - , , , (9)
- , , , , , , , (10)
$ , (11)
- , , , , , , (12)
& , (13)
, , , ,
, , ,
, , , (14)1
, , , (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
, - (2)
0 , (3) 2 (4)
. , (5)
9: # 56;,5/65/6 5 , (6)
2 2 , (7)
. , (8) - , , , (9)
- , , , , , , , (10)
$ , (11)
- , , , , , , (12)
& , (13)
, , , ,
, , ,
, , , (14)1
, , , (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
, - (2)
0 , (3) 2 (4)
. , (5)
9 , 5/656 #
<<=>6, (6)
2 2 , (7)
. , (8) - , , , (9)
- , , , , , , , (10)
$ , (11)
- , , , , , , (12)
& , (13)
, , , ,
, , ,
, , , (14)1
, , , (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
, - (2)
0 , (3) 2 (4)
. , (5)
9 , 5 , (6)
25656 2 , (7)
. , (8) - , , , (9)
- , , , , , , , (10)
$ , (11)
- , , , , , , (12)
& , (13)
, , , ,
, , ,
, , , (14)1
, , , (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
, - (2)
0 , (3) 2 (4)
. , (5)
9 , 5 , (6)
2 25656 # ,%
>6<=>6 (7)
. , (8) - , , , (9)
- , , , , , , , (10)
$ , (11)
- , , , , , , (12)
& , (13)
, , , ,
, , ,
, , , (14)1
, , , (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
, - (2)
0 , (3) 2 (4)
. , (5)
9 , 5 , (6)
2 2 , (7)
./ # . ' -. # 7, 8 % 01 >6<=>6 (8) - , , , (9)
- , , , , , , , (10)
$ , (11)
- , , , , , , (12)
& , (13)
, , , ,
, , ,
, , , (14)1
, , , (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
, - (2)
0 , (3) 2 (4)
. , (5)
9 , 5 , (6)
2 2 , (7)
. , (8) -. # %,7,01 >6<=>6, (9)
- , , , , , , , (10)
$ , (11)
- , , , , , , (12)
& , (13)
, , , ,
, , ,
, , , (14)1
, , , (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
, - . (2)
0 , (3) 2 (4)
. , (5)
9 , 5 , (6)
2 2 , (7)
. , (8) - , , , (9)
-$ # %,?1,7,01 >6<=>6 # ,%,$,01
>6<=>6, (10)
$ , (11)
- , , , , , , (12)
& , (13)
, , , ,
# ,$ % & % $,01 ,
# ,B( % $,01 , (14)1
, , , (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
, - . (2)
0 , (3) 2 (4)
. , (5)
9 , 5 , (6)
2 2 , (7)
. , (8) - , , , (9)
- , , , , , , , (10)
$/ # $ ' -$ # $, 8 % 01 >6<=>6 (11)
- , , , , , , (12)
& , (13)
, , , ,
, , ,
, , , (14)1
, , , (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
, - . (2)
0 , (3) 2 (4)
. , (5)
9 , 5 , (6)
2 2 , (7)
. , (8) - , , , (9)
- , , , , , , , (10)
$ , (11)
-& # %,?@,A,0@ ><=> ,# %,&,0@ >
<=> (12)
& , (13)
, , , ,
, , ,
, , , (14)1
, , , (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
, - . (2)
0 , (3) 2 (4)
. , (5)
9 , 5 , (6)
2 2 , (7)
. , (8) - , , , (9)
- , , , , , , , (10)
$ , (11)
- , , , , , , (12)
&/ # &' -& # &, 8 % 0@ ><=> (13)
, , , ,
# ,$ % & % $,01 ,
# ,B( % $,01 , (14)1
, , , (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
, - (2)
0 , (3) 2 (4)
. , (5)
9 , 5 , (6)
2 2 , (7)
. , (8) - , , , (9)
- , , , , , , , (10)
$ , (11)
- , , , , , , (12)
& , (13)
B(/ # ,$, 8 % 019:
8 ' 9: % ,&, 8 % 0@9
8 ' 9
, , ,
, , , (14)1
, , , (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
, - (2)
0 , (3) 2 (4)
. , (5)
9 , 5 , (6)
2 2 , (7)
. , (8) - , , , (9)
- , , , , , , , (10)
$ , (11)
- , , , , , , (12)
& , (13)
, , , ,
# ,$ % & % $,01 >6<=>6 ' &,0@
><=>
, , , (14)1
, , , (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
, - . (2)
0 , (3) 2 (4)
. , (5)
9 , 5 , (6)
2 2 , (7)
. , (8) - , , , (9)
- , , , , , , , (10)
$ , (11)
- , , , , , , (12)
& , (13)
, , , ,
, , ,
# ,B( % $,01 >6<=>6 ' &,0@
><=> (14)1
, , , (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
, - (2)
0 , (3) 2 (4)
. , (5)
9 , 5 , (6)
2 2 , (7)
. , (8) - , , , (9)
- , , , , , , , (10)
$ , (11)
- , , , , , , (12)
& , (13)
, , , ,
# ,$ % & % $,01 ,
# ,B( % $,01 , (14)1
!"/ # B(/ ' ()* ' (+* or
, , , (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
, - . (2)
0 , (3) 2 (4)
. , (5)
9 , 5 , (6)
2 2 , (7)
. , (8) - , , , (9)
- , , , , , , , (10)
$ , (11)
- , , , , , , (12)
& , (13)
, , , ,
# ,$ % & % $,01 ,
# ,B( % $,01 , (14)1
!"/ # !" % $,01 >6<=>6 ' &,0@
><=>, (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
, - (2)
0 , (3) 2 (4)
. , (5)
9 , 5 , (6)
2 2 , (7)
. , (8) - , , , (9)
- , , , , , , , (10)
$ , (11)
- , , , , , , (12)
& , (13)
, , , ,
, , ,
, , , (14)1
, , , (15)
CD/5EFG #
CD5EFG %
H5EFG 01
>6<=>6 , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
, - (2)
0 , (3) 2 (4)
. , (5)
9 , 5 , (6)
2 2 , (7)
. , (8) - , , , (9)
- , , , , , , , (10)
$ , (11)
- , , , , , , (12)
& , (13)
, , , ,
, , ,
, , , (14)1
, , , (15)
C ' I5EFG ,0@
><=> (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B , , , ,
12 In the special case that θX = θM = θ, then
, - (2)
0 , (3) 2 (4)
. , (5)
9 , 5 , (6)
2 2 , (7)
. , (8) - , , , (9)
- , , , , , , , (10)
$ , (11)
- , , , , , , (12)
& , (13)
, , , ,
, , ,
, , , (14)1
, , , (15)
C , (16)!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 B(/ # ,B( % 0, $, >6
<=>6 % &,>
<=> .
Therefore, the cyclically adjusted currentaccount balance (as a percentage of GDP)results from the difference between the actualbalance and the cyclical component. The cycli-cal component, in turn, depends on the shareof nominal exports and imports in GDP, theincome elasticities of exports and imports andfinally the output gaps of the home economyand of its trading partners.
Thus, in order to isolate the cyclical part of theadjustment, two elements of information arerequired. The first is related to the intensity ofthe economic cycle in the home economy andits major trading partners, which is measuredby the magnitude of their output gaps. The sec-ond concerns the sensitivity of the currentaccount balance to changes in the output gap,which relates to the export and import elas-ticity of GDP of the trading partners and thehome country, respectively.
The impact of the economic cycle on the exter-nal sector is primarily reflected in its impact onthe trade balance, which is the main part of thecurrent account balance. Changes in the othertwo parts of the current account (primary andsecondary income balances) are considered tobe structural in nature.13
In our research, we do not estimate the incomeelasticities of export and imports; we ratheremploy the elasticities estimated in otherresearch work. In our baseline scenario, weemploy an elasticity of 1.5 for both exports andimports with respect to income, which reflectsthe European Commission’s approach asdescribed in Salto and Turrini (2010). We alsoneed to assume home imports and exports asisoelastic to home and foreign GDP, respec-tively. This means that they are exogenouslygiven constant long-run elasticities, θX and θM
(Fabiani et al. 2016; Amador and Silva 2019).
The robustness of our results is confirmed bya sensitivity analysis based on a wide range ofelasticities from 1.0 to 2.0. This approachallows us to identify the potential discrepanciesamong research work that employed different
import and export income elasticities. Finally,as trade elasticities differ over time (let aloneamong countries), our sensitivity analysisallows us to better assess the cyclical and non-cyclical adjustment of the Greek currentaccount deficit.
For the current account components, the Bankof Greece Balance of Payments data for theperiod 2002-2019 were used. Greece’s outputgap, as estimated by the Bank of Greece, wasemployed in our analysis (Bank of Greece2018b and 2019b), while trading partners’ out-put gap is the trade-weighted output gap of themain trading partners, as provided byAMECO and the IMF’s World Economic Out-look, and using the ECB’s trade shares (seeAppendix A).
4 ANALYSIS AND RESULTS
In the period 2003-2010, Greece had a positiveoutput gap, which indicated that actual importsin that period were above the level that corre-sponds to potential output. Starting from 2011,Greece moved into a negative output gap ―asa result of the Greek economic crisis― whichpeaked in 2013; thereafter, the output gap hasbeen following a gradually narrowing trend(see Chart 2). Thus, imports have been belowthe level that corresponds to potential output.
Turning to Greece’s main trading partners, themagnitude of their trade-weighted output gapis significantly smaller, compared with Greece.This reflects, on the one hand, the magnitudeof the Greek recession and, on the other hand,the fact that trading partners’ output gap is cal-culated as the weighted average of 32 countriesthat may stand at different points in their eco-nomic cycles. Overall, for the period 2009-2016, trading partners were ―on average―faced by a mildly negative output gap, partly
51Economic BulletinJuly 202080
13 After all, most of these categories include, on the one hand,receipts, in particular inflows from EU structural funds, and, on theother hand, outflows, primarily interest payments to the supportmechanism for servicing public debt. Therefore, they are not linkedwith the economic cycle, as they are structural in nature.
reflecting the global financial crisis and theeuro area sovereign crisis (see Chart 2). There-fore, Greek exports were below the level thatis consistent with the potential output of trad-ing partners in that period.
As discussed in Section 2, the actual currentaccount deficit kept increasing from 2005 until2008, as both the cyclical and the non-cyclicalcomponents of the trade balance were deteri-orating; the former due to the positive outputgaps in Greece and in its trading partners.Over the same period, the balance of primaryand secondary income (bpi+bsi) was in deficitand widening, before stabilising at around -3.3% of GDP in 2007. From 2009 onwards,the current account deficit is on a downwardtrajectory and stabilised at around 2% of GDP―on average― in the 2016-2019 period. In the
first years of the Greek economic crisis (2011-2012), the actual current account adjustmentwas mainly driven by the cyclical component(i.e. Greece’s and its trading partners’ outputgaps) and secondarily by the balance of pri-mary and secondary income. In the years thatfollowed (2013-2015), the non-cyclical com-ponent of the trade balance came into play andsupported a further adjustment in the actualcurrent account. In the 2016-2019 period, thecyclically-adjusted trade balance hoveredaround -3.7% of GDP and the respective cur-rent account stood at -4.5% of GDP. The cycli-cal component gradually decreased andremained ―on average― at 2.6% of GDP.
The highest current account deficit wasrecorded in 2008, amounting to EUR 36.6 bil-lion, i.e. 15.1% of GDP (see Chart 1). The
51Economic Bulletin
July 2020 81
51Economic BulletinJuly 202082
cyclical part of this deficit was 3.3% of GDP,with the consequent cyclically adjusted deficitof the current account reaching 11.8% of GDP(see Chart 3). In 2019, the current accountdeficit fell to 1.4% of GDP, recording animprovement of 13.7 pp of GDP since 2008. Ofthis improvement, 5.4 pp of GDP is attributedto cyclical factors and 8.3 pp to structural fac-tors (see Chart 4). Therefore, 60% of the cur-rent account deficit correction can be attrib-uted to structural factors, and the remaining40% to cyclical factors (trading partners’ andGreece’s output gap). In greater detail, theimprovement in the trade balance (goods andservices) was 10.9 pp of GDP, almost equallydistributed between the cyclical and the non-cyclical component. The improvement in theprimary and secondary income balance,which by definition is structural, amounted to2.8 pp of GDP.14
At this point, it would be useful to see how ourresults compare with the corresponding ones
of previous studies. First, Tressel and Wang(2014) estimated the current account adjust-ment decomposition for the 2007-2012 periodfor a number of euro area countries. In thecase of Greece, cyclical factors (including out-put gap, financial conditions and commodityterms of trade) were estimated to have con-tributed 50% of the actual current accountreversal during the above mentioned period.
The ECB (2014) estimated that, in the case ofGreece, more than 50% of the current accountadjustment in the 2008-2012 period wasexplained by cyclical factors. This outcome wasbased on the estimated output gaps by theIMF, the European Commission (EC) and theOECD and followed the first stream method-ology described briefly in Section 3. Withoutprejudice to data revision (both currentaccount and output gap) since 2014, the ECB’s
14 Even if the EU’s AMECO database estimates for Greece’s outputgap are applied, the results are not significantly different (36%cyclical and 64% structural).
51Economic Bulletin
July 2020 83
conclusion is in line with our analysis, whichindicates that more than 60% of the cumula-tive adjustment could be attributed to cyclicalfactors in that period. The EC (2015) also sup-ports that in the initial phase of the adjustmentin the deficit countries, the current accountdeficit contraction was associated withreduced imports on the back of weakeneddomestic demand.
More recently, the IMF (2019) attempted todecompose the Greek current account adjust-ment into cyclical and structural factors byusing the first stream methodology and utilis-ing ―to some extent― the IMF’s EBAmethodology. They considered as cyclical theoutput gap, the commodity terms of trade, thecredit-to-GDP gaps and volatility (proxied bythe VIX index), while structural factorsincluded variables such as the cyclicallyadjusted fiscal balance, real unit labour cost(ULC), measures of the institutional and polit-ical environment, etc. They concluded that the
largest part of the deterioration in the currentaccount between 2004 and 2008 was structural.For the 2008-2018 period, about 75% ofGreece’s current account correction was attrib-uted to cyclical factors and the remaining 25%to structural factors and policies.
Our analysis, though, indicates that more than50% of the cumulative adjustment for the2008-2018 period can be attributed to non-cyclical factors. This discrepancy, apart frommethodological differences, may partly reflectdata revisions, both of current account dataand of output gaps. It should also be notedthat in Cubeddu et al. (2019) regarding theEBA 2018 update, the credit gap (i.e. thecredit-to-GDP gap) is considered a policyvariable rather than a cyclical factor as in theabovementioned research. Had the credit-to-GDP gap been included in the structural fac-tors, the breakdown would have been differ-ent and definitely in favour of the structuraladjustment.
5 SENSITIVITY ANALYSIS
In order to investigate the sensitivity of theresults to the income elasticities of imports andexports, the current account balance adjustmentwas calculated for a range of elasticities (1.0-2.0) around the central value (1.5). The valuesof the elasticities are in line with a number ofprevious research endeavours (see AppendixB). The effect of the income elasticity ofimports is greater than that of exports. Struc-tural factors may account for 47%-74% of thecurrent account deficit correction, dependingon the elasticities used (see the table above).
Overall, as regards the trade balance, much ofthe improvement came from the significantrecovery in exports, particularly after 2012,which, coupled with the almost zero output gapof trading partners’ economies after 2016, waspurely structural in nature.15 On the otherhand, the decline in imports was primarilycyclical, as a result of the decline in domesticdemand and the emergence of a large negativeoutput gap. However, to the extent that theprolonged recession also contributed to adecline in potential output, it helped reducethe cyclical portion of the overall adjustment.Finally, given the high import content ofexports, the increase in exports also con-tributed to a rise in the structural share ofimports.16 More generally, the structuralchanges that took place in recent years in theGreek economy and the consequent shift,albeit gradual, towards a more extrovertedgrowth model appear to be reflected in the sig-
nificant structural adjustment of the currentaccount balance.17
6 CONCLUSIONS
Greece’s external imbalances widened per-sistently in the lead-up to the 2008 financialcrisis and have narrowed considerably after-wards, as the economy went through a majorrecession. The assessment of the relativeimportance of structural and cyclical factorsin explaining the correction of currentaccount balances is key in evaluating thefuture development of the external sector. Inthis paper we assessed the relative signifi-cance of structural and cyclical factors behindthe considerable reduction of the Greek cur-rent account deficit and the analysis reliedprimarily on the methodology used by Halt-maier (2014). In particular, the cyclical partof the adjustment was estimated using alter-native income elasticities of both imports andexports in line with the literature, while thestructural part was calculated indirectly as thedifference between the overall actual balanceand the cyclical part.
Our analysis shows that the current accountadjustment has in a considerable part been
51Economic BulletinJuly 202084
15 See for instance Bardaka and Papazoglou (2019) and Backinezoset al. (2019).
16 For the participation of Greece in global value chains, see forinstance Gibson et al. (2019).
17 According to Bank of Greece estimates, over the 2010-2017 period(data availability), the volume of tradable goods and services recordeda cumulative increase of 14%, compared with non-tradables.
1.0 26 74 27 73 27 73
1.5 39 61 40 60 40 60
2.0 52 48 52 48 53 47
Income elasticity of imports (θM)
Income elasticity of exports (θX)
1.0 1.5 2.0
cyclical non-cyclical cyclical non-cyclical cyclical non-cyclical
Sensitvity analysis with respect to income elasticity of exports and imports
(cumulative share of cyclical and non-cyclical adjustment in the current account deficit, 2008-2019)
Source: Authors' calculations.
driven by structural factors, which account for60% of the overall adjustment according toour baseline scenario. Therefore, most of thecorrection is of a permanent nature, whichmeans that the relatively high deficits of thepre-crisis period are not expected to re-emerge any time soon. Our results are robustas they largely hold under alternativehypotheses regarding the income elasticitiesof exports and imports. That is, even thoughthe deep recession that the country experi-enced has indeed contributed to the correc-tion of the external imbalance, most of theadjustment was structural, reflecting a grad-ual shift in Greece’ s growth model. This hasmanifested itself to a significant extent in thefact that the country is turning into a more
open economy as a result of its enhancedexport performance.
However, given the fact that a significant partof the adjustment came from cyclical factorsand the high negative size of net internationalinvestment position, the above conclusionsshould not lead to complacency. On the con-trary, persistent structural reforms aimed atenhancing the overall competitiveness of theGreek economy and shifting productiontowards tradable goods and services shouldremain among the main economic policy pri-orities. That is, targeted policy initiatives mustbe at the centre of the reform effort, as theyare essential in sustaining the reduction ofexternal imbalances over the medium term.
51Economic Bulletin
July 2020 85
Algieri, B. (2014), “Drivers of export demand: a focus on the GIIPS countries”, The World Econ-omy, 37(10), 1454-1482, October.
Amador, J. and J.F. Silva (2019), “Cyclically-adjusted current account balances in Portugal”,Banco de Portugal Economic Studies, V(1), Banco de Portugal, Lisbon, January.
Athanasoglou, P.P., C. Backinezos and E.A. Georgiou (2010), “Export performance and com-modity composition”, Bank of Greece Working Paper No. 114.
Backinezos, C., S. Panagiotou and A. Rentifi (2019), “Greek export performance: a constant mar-ket share analysis”, Bank of Greece, Economic Bulletin, 50, December.
Bank of Greece (2017), Annual Report 2016, Box V.3 “Export performance and import pene-tration”, February.
Bank of Greece (2018a), Monetary Policy 2017-2018, Box IV.2 “The import content of aggre-gate demand in Greece”, July.
Bank of Greece (2018b), Monetary Policy Interim Report 2018, Box IV.1 “The key drivers oflong-term growth for the Greek economy”, December.
Bank of Greece (2019a), Monetary Policy Interim Report 2019, Box IV.3 “Developments andprospects of inbound tourism (2010-2019)”, July.
Bank of Greece (2019b), “Potential output methodology”, Economic Analysis and ResearchDepartment, Econometric Forecasting Section, mimeo.
Bank of Greece (2020), Annual Report 2019, Box IV.4 “Current account developments from thesaving-investment perspective”, March.
Bardaka, I. (2017), “Trade equations for Greece”, Bank of Greece, mimeo.Bardaka, I. and C. Papazoglou (2019), “The determinants of Greece’s export supply of oil”, Bank
of Greece, Economic Bulletin, 49, JulyBragoudakis, Z., D. Sideris, P. Petroulas and E. Vourvachaki (2016), “The annual econometric
model of the Bank of Greece”, Bank of Greece, mimeo.Caporale, G.M. and M.K.F. Chui (1999), “Estimating income and price elasticities of trade in
a cointergation framework”, Review of International Economics, 7(2), 254-264. Christodoulopoulou, S. and O. Tkačevs (2016), “Measuring the effectiveness of cost and price
competitiveness in external rebalancing of euro area countries: What do alternative HCIs tellus?”, Empirica, 43, 487-531.
Cubeddu, L., S. Krogstrup, G. Adler, P. Rabanal, M.C. Dao, S.A. Hannan, L. Juvenal, N. Li, C.O.Buitron, C. Rebillard, D. Garcia-Macia, C. Jones, J. Rodriguez, K.S. Chang, D. Gautam andZ. Wang (2019), “The external balance assessment methodology: 2018 update”, IMF WorkingPaper No. 65, International Monetary Fund, March.
ECB (2014), “To what extent has the current account adjustment in the stressed euro area coun-tries been cyclical or structural?”, Box 5, European Central Bank, Monthly Bulletin, January.
European Commission (2015), “Rebalancing in the euro area: an update”, Box 1.3, EuropeanEconomic Forecast, Spring 2015.
Fabiani, S., S. Federico and A. Felettigh (2016), “Adjusting the external adjustment: cyclical fac-tors and the Italian current account”, Questioni di Economia e Finanza (Occasional Papers),No. 346, Bank of Italy.
Greek Shipping Co-operation Committee (various years), “Greek controlled shipping”. Gibson, H.D., G. Pavlou, C. Tsochatzi and M. Vasardani (2019), “Greece’s integration into global
value chains”, Bank of Greece, Economic Bulletin, 50, December.Haltmaier, J. (2014), “Cyclically adjusted current account balances”, International Finance Dis-
cussion Paper No. 1126, Board of Governors of the Federal System.Hubrich, K. and T. Karlsson (2010), “Trade consistency in the context of the Eurosystem pro-
jection exercises: an overview”, ECB Occasional Paper No. 108, European Central Bank,March.
51Economic BulletinJuly 202086
R E F E R ENC E S
IMF (2009), Balance of payments and international investment position manual, Sixth Edition(BPM6), Washington, D.C.: International Monetary Fund.
IMF (2019), “The quality of Greece’s external adjustment”, Greece: Selected Issues, IMF Coun-try Report No. 19/341, 26-28, October.
Morin, M. and C. Schwellnus (2014), “An update of the OECD international trade equations”,OECD Economics Department Working Paper No. 1129, OECD Publishing, Paris.
Pain, N., A. Mourougane, F. Sedillot and L. Le Fouler (2005), “The new OECD internationaltrade model”, OECD Economics Department Working Paper No. 440, OECD Publishing, Paris.
Papazoglou, C. (2007), “Greece’s potential trade flows: a gravity model approach”, InternationalAdvances in Economic Research, 13(4), 403.
Phillips, S., L. Catão, L. Ricci, R. Bems, M. Das, J. Di Giovanni, D.F. Unsal, M. Castillo, J. Lee,J. Rodriguez and M. Vargas (2013), “The External Balance Assessment (EBA) Methodology”,IMF Working Paper No. 272, International Monetary Fund, December.
Salto, M. and A. Turrini (2010), “Comparing alternative methodologies for real exchange rateassessment”, European Commission, Directorate-General for Economic and Financial Affairs,European Economy, Economic Papers, No. 427, September.
Senhadji, A.S. and C.E. Montenegro (1999), “Time series analysis of export demand equations:a cross-country analysis”, IMF Staff Papers, 46(3), September/December 1999.
Sideris, D. and N.G. Zonzilos (2005), “The Greek model of the European System of CentralBanks Multi-Country Model”, Bank of Greece Working Paper No. 20.
Tressel, T. and S. Wang (2014), “Rebalancing in the euro area and cyclicality of current accountadjustments”, IMF Working Paper No. 14/130, July.
51Economic Bulletin
July 2020 87
Trading partners’ GDP (potential GDP) wasconstructed as their trade-weighted GDP(potential GDP). We have used the exporttrade weights of Greece’s main trading part-ners in manufactured goods, as defined in Sec-tions 5 to 8 of the Standard InternationalTrade Classification (SITC 5-8), i.e. excludingtrade in services, agricultural products, rawmaterials and energy products. As data wereavailable up until 2015, for the 2016-2019period the trade weights refer to the last avail-able observation (i.e. 2013-2015 period level).The trade weights, as calculated by the ECB,are based on bilateral data on trade in manu-factured goods for the periods 1995-97, 1998-2000, 2001-03, 2004-06, 2007-09, 2010-12 and2013-15, the last update having been made in2017. The export trade weights were retrievedfrom the ECB Statistical Data Warehouse(key: WTS.A.GR.xx.Z0Z.X0.X.MAN.F). Theaggregated trade share of the 32 countries usedamounts to approximately 70% of Greekexports of manufactured goods (see Table A1).The selection of these countries was alsodriven by the availability of potential outputdata in AMECO or the IMF’s WEO. The
methodology for the construction of the geo-metric weighted average of trading partners’GDP and potential GDP is similar to the oneof the export demand index in Hubrich andKarlsson (2010).
GDP data were retrieved from the Eurostatwebsite. AMECO potential GDP data(updated on 6 May 2020) were used, exceptfor Australia, Canada, Japan and Norway.For these four countries, the IMF’s October2019 World Economic Outlook database wasused, as the respective data were not avail-able in AMECO. It should be noted that theApril 2020 WEO database does not includepotential output data, due to the high level ofuncertainty in current global economic con-ditions.
Finally, trading partners’ GDP and potentialGDP, reflecting foreign demand and potentialforeign demand, respectively, were chain-linked so as to avoid any changes stemmingfrom the variation of the trade weights acrossperiods (i.e. 1995-97, 1998-2000, 2001-03,2004-06, 2007-09, 2010-12 and 2013-15).
51Economic BulletinJuly 202088
APP END I X ACON S T RUC T I ON O F T R ADE -WE I GH T ED T R AD I NG PAR TNER S ’ GDP AND PO T ENT I A L GDP
Austria Bulgaria Australia
Belgium Croatia Canada
Cyprus Czechia Japan
Estonia Denmark Norway
Finland Hungary United Kingdom
France Poland United States
Germany Romania
Ireland Sweden
Italy
Latvia
Lithuania
Luxembourg
Malta
Netherlands
Portugal
Slovakia
Slovenia
Spain
Euro area EU – non-euro area Non-EU
Table A1 Trading partners
51Economic Bulletin
July 2020 89
Senhadji andMontenegro (1999)
2.81 (OLS and Phillips-Hansen’s Fully Modified)
1960-1993Real exports of goodsand non-factor services
Weighted average oftrading partners’ GDPminus exports
Caporale and Chui(1999)
2.38 (DOLS); 2.10 (ARDL)
1960-1992 (annual data)
Export volumes of goodsand services
Export shares-weightedGDP
Pain et al. (2005)
A unit income elasticityimposed
1982-2002 (quarterly data)
Export volumes of goodsand services
Export market share(market share-weightedsum of partner countries’imports)
Sideris and Zonzilos(2005)
1.00 (supported by data)
1980:q1-2000:q4 (quarterly data)
Export volumes of goodsand services
Weighted average ofimport volumes of maintrading partners
Algieri (2014)
1.81(range: 1.41-2.46)1
(VECM)
1980:q1-2012:q3 (quarterly data)
Real exports of goodsand services
– Corrected world GDP:real world GDP in USDadjusted for PPP
– Foreign demand:weighted average ofimport volumes of maintrading partners
Morin and Schwellnus(2014)
1.00 (imposed, but supportedby data)
1992:q2-2012:q2 (quarterly data)
Export volumes of goodsand services
Export market share(market share-weightedsum of partner countries’imports)
Christodoulopoulou and Tkačevs (2016)
0.91 (range: 0.43-1.45)2
2000:q1-2013:q1 (quarterly data)
Export volumes of goods Geometric weightedaverage of tradingpartners’ imports
0.93 (range: 0.87-0.98)2
Export volumes ofservices
Bragoudakis et al. (2016)
0.83 (Engle-Grangercointegration)
1980-2015 (annual data)
Export volumes of goodsand services
Weighted average ofimport volumes of maintrading partners
Bardaka (2017)
1.60 (Johansencointegration)
2002-2017 (quarterly data)
Export volumes of goodsand services
Trade share-weightedpartner countries’ GDP
Researcher Foreign income elasticity Estimation period Dependent variable Foreign income
Table B1 Exports
A P P END I X BLONG - RUN E S T IMA T I ON S F OR I N COMEE L A S T I C I T I E S O F E XPOR T S AND IMPORT S
1 The figure refers to the average estimated long-run elasticity of three alternative models. The foreign income elasticity was estimated at 2.46when corrected world GDP was used and at 1.56 and 1.41 when foreign demand was used.2 The figure refers to the average estimated long-run elasticities of a model with different Harmonised Competitiveness Indicators (HCI) basedon CPI, CPI for Services, PPI, ULCM, ULCT and GDP deflators. Significant at the 1% and 5% level.
51Economic BulletinJuly 202090
Caporale and Chui(1999)
1.32 (DOLS); 1.58 (ARDL)
1960-1992 (annual data)
Import volumes of goodsand services
Real GDP
Pain et al. (2005)
A unit income elasticityimposed
1982-2002 (quarterly data)
Import volumes of goodsand services
Components of domesticdemand
Sideris and Zonzilos(2005)
1.00 (imposed)
1980:q1-2000:q4(quarterly data)
Import volumes of goodsand services
Weighted sum of theimport content of thedomestic demandcomponents
Morin and Schwellnus(2014)
1.621 1992:q2-2012:q2(quarterly data)
Import volumes of goodsand services
Components of domesticdemand
Christodoulopoulou and Tkačevs (2016)
1.56 (range: 1.47-1.68)2
2000:q-2013:q1 (quarterly data)
Import volumes of goods Private consumption + gross capital formation + governmentconsumption
0.22 (range: -0.07-0.57)3
Import volumes ofservices
Bragoudakis et al. (2016)
1.0 (restricted); 1.23 (unrestricted)(Engle-Grangercointegration)
1980-2015 (annual data)
Import volumes of goodsand services
Weighted sum of theimport content of thedomestic demandcomponents
Bardaka (2017)
1.41 (Johansencointegration)
2002-2017 (quarterly data)
Import volumes of goodsand services
Real GDP
ResearcherDomestic income elasticity Estimation period Dependent variable Domestic income
Table B2 Imports
1 Based on the estimated impulse response for import volume generated by a 1% increase in total expenditure after 70 quarters (see Table A.7therein).2 The figure refers to the average estimated long-run elasticity of a model with different Harmonised Competitiveness Indicators (HCI) basedon CPI, CPI for Services, PPI, ULCM, ULCT and GDP deflators. Significant at the 1% level.3 The figure refers to the average estimated long-run elasticity of a model with different Harmonised Competitiveness Indicators (HCI) basedon CPI, CPI for Services, ULCT and GDP deflators. Not significant at the 10% level.
Yorgos KorfiatisPayment and Settlement Systems Department
ABSTRACTSince 2008 new schemes such as digital currencies have emerged in limited forms. Currently, dig-ital technologies enjoy a widespread use in the area of payments and, as a result, central bankshave started contemplating the costs and benefits of issuing a digital currency for broad use inpayments and settlements. This paper defines Central Bank Digital Currency (CBDC) and clas-sifies it alongside existing forms of money. Further, it highlights the properties that such a cur-rency should have, and analyses the main motivations behind the potential issuance of an account-based retail CBDC (called here the “d-euro”). We argue that the latter should involve a public-private partnership whose effectiveness will be enhanced by the establishment of separate anddistinct roles. This paper briefly discusses legal considerations, monetary policy implications andfinancial stability issues arising from d-euro. In doing so, it further analyses d-euro’s underly-ing principles, which would at the same time constitute its key benefits. We provide a generalframework and propose the technical design of our so called “two-dot model”. Issues concern-ing parity, supply limits and anonymity are dealt with. The proposed technical design seems aneffective and workable solution that boosts the autonomy and resilience of the European pay-ment systems. Finally, we argue that the initial introduction of d-euro as a retail payment mediummay establish it as the new cash of a digitalised ecosystem, reflecting a new digital trust issuedby central banks.
Keywords: CBDC; d-euro; money; Dedicated Digital Account; TIPS
JEL classification: E58; G21; O33
51Economic Bulletin
July 2020 91
D - EURO : I S S U I NG TH E D I G I T A L T RU S T
D - EURO : ΕΚΔ Ι ΔΟΝΤΑΣ ΤΗΝ ΨΗΦΙΑΚΗ Π Ι Σ ΤΗ
Γιώργος ΚορφιάτηςΔιεύθυνση Συστημάτων Πληρωμών και Διακανονισμού
ΠΕΡΙΛΗΨΗΑπό το 2008 περίπου τα ψηφιακά νομίσματα έχουν εμφανιστεί με διάφορες μορφές. Σήμερα,καθώς οι ψηφιακές τεχνολογίες χρησιμοποιούνται ευρέως στο χώρο των πληρωμών, οι κεντρι-κές τράπεζες έχουν αρχίσει να μελετούν τα κόστη και τα οφέλη της έκδοσης ενός ψηφιακού νομί-σματος με ευρεία χρήση στις πληρωμές και το διακανονισμό. Το παρόν άρθρο δίνει τον ορι-σμό ενός τέτοιου ψηφιακού νομίσματος κεντρικής τράπεζας (CBDC) και το ταξινομεί σε σχέσημε τις υπάρχουσες μορφές χρήματος. Στη συνέχεια, διευκρινίζει τις ιδιότητες που πρέπει να δια-θέτει και εξηγεί τα κύρια αίτια για τη δυνητική έκδοση ενός CBDC το οποίο προτείνεται ναβασίζεται σε τραπεζικό λογαριασμό (και θα ονομάζεται d-euro). Η προσπάθεια αυτή απαιτείτη σύμπραξη δημόσιου και ιδιωτικού τομέα, η αποτελεσματικότητα της οποίας θα ενισχυθεί απότον καθορισμό διακριτών ρόλων για τον κάθε τομέα. Το άρθρο παρουσιάζει συνοπτικά τουςσχετικούς νομικούς προβληματισμούς, καθώς και τις δυνητικές επιπτώσεις μιας ενδεχόμενηςέκδοσης του d-euro σε θέματα χρηματοπιστωτικής σταθερότητας και νομισματικής πολιτικής.Επίσης, αναλύονται οι βασικές αρχές του νομίσματος, οι οποίες συγχρόνως θα αποτελούν ταβασικά του πλεονεκτήματα. Προτάσσεται ένα γενικό πλαίσιο λειτουργίας που συνοδεύεται απότο τεχνικό σχέδιο του αποκαλούμενου “μοντέλου δύο σημείων”. Στη συνέχεια, επεξηγούνταιθέματα σχετικά με την ισοτιμία, την προσφορά χρήματος και την ανωνυμία. Ο προτεινόμενοςτεχνικός σχεδιασμός κρίνεται ως μια κατάλληλη και εφαρμόσιμη λύση, ικανή να ενισχύσει τηναυτονομία και την ανθεκτικότητα των ευρωπαϊκών συστημάτων πληρωμών. Τέλος, υποστηρί-ζεται ότι μια ενδεχόμενη αρχική υιοθέτηση του d-euro ως μέσου πληρωμών λιανικής θα μπο-ρούσε να οδηγήσει στην καθιέρωσή του ως της νέας μορφής μετρητών ενός ψηφιακού οικοσυ-στήματος, αντανακλώντας μια νέα ψηφιακή πίστη προερχόμενη από τις κεντρικές τράπεζες.
51Economic BulletinJuly 202092
1 INTRODUCTION
A few years ago one frequently heard that theuse of physical cash would eventually decline(Cœuré 2018). A decade ago one read linessuggesting that “we have proposed a system forelectronic transactions without relying ontrust” (Nakamoto 2008, p. 8). Nowadays, it iscommon to see someone pay at a supermarketby swiping a card or by waving their mobilephone, irrespective of the currency or theamount to be paid. In the near future, one maydecide to pay for a pizza by using stablecoins,i.e. cryptoassets1 with value stabilising charac-teristics. Such assets, mainly used throughwidespread social messaging apps, could makepayments cheaper and faster. However, theyhave to overcome regulatory and supervisorychallenges at the international policy level,especially regarding risks to financial stabilityand monetary sovereignty, data protection andcompliance with anti-money laundering andcounter-terrorism financing rules (G7 Work-ing Group on Stablecoins 2019; EuropeanCentral Bank 2019c; FATF 2020).
This paper does not intend to rake over thecoals of the stablecoins issue once more. It sim-ply uses this new type of cryptocurrency, whichapplies stabilisation mechanisms in order tominimise price volatility, as an introductorypoint of my discourse about digital currencies(see for example Bullmann et al. 2019). It is astarting point for central banks in their effortsto use new technologies in order to enhancethe efficiency and resilience of the paymentslandscape by issuing to the public an instant,safe, cheap, and potentially semi-anonymousdigital currency, i.e. a retail Central Bank Dig-ital Currency (CBDC)2.
We start by mentioning a set of truisms. Dig-ital currencies have existed for a number ofdecades. David Chaum (1983), who is widelyrecognised as the inventor of digital cash,
introduced the cryptographic primitive of ablind signature that could prevent doublespending. In 2008, Bitcoin was the first cryp-tocurrency that used permissioned blockchainprotocols, relying heavily on cryptography, thusaccomplishing nearly anonymous transactionsthrough decentralised means. The key inno-vation in this regard is the introduction of a“distributed ledger”, which allows a digital cur-rency to be used in a decentralised paymentsystem (Böhme et al. 2015; Narayanan et al.2017). A common feature of developedeconomies has been the wide variety of ways inwhich payments are made as well as the manyforms of money used. There is also a widerange of economic agents whose liabilities canfunction as money. The most prominent issuersof money are central banks, which provide cen-tral bank money in the form of both banknotesand liquidity through the financial system, andcommercial banks, which generally issue pri-vate money (commercial bank money) in theform of deposits (Bank for International Set-tlements 2003; Kokkola 2010). Contrary to thevarious decentralised forms of assets, centralbanks also offer a form of digital currency tocertain institutions, the so-called reserve bal-ances, i.e. deposits held with the central bank(Friedman 2000; Agarwal and Kimball 2019).On the other side, commercial banks offer tothe public deposits that can be converted tocash (legal tender) through withdrawal (Mid-dleton and Sinha 2019).
In this context, the European Central Bank(ECB) does recognise that, in today’s contin-uously changing payments ecosystem, greatattention should be given to the front-end ofEuropean retail payments systems, and in par-
51Economic Bulletin
July 2020 93
D - EURO : I S S U I NG TH E D I G I T A L T RU S T
Yorgos Korfiatis*
Payment and Settlement Systems Department
* The author would like to thank Stelios Chrysomallis for his usefulcontribution and Eleftheria Kostika for a fruitful discussion. Theviews expressed are those of the author and do not necessary reflectthose of the Bank of Greece. Any errors and omissions are theauthor’s responsibility.
1 See Chimienti et al. 2019 and European Central Bank 2019a.2 CBDC or d-euro are here used interchangeably.
ticular to the point-of-sale and online paymentsolutions (Cœuré 2019c). The necessity for aretail payments strategy has been the result ofthree main reasons. The first relates to thestrategic autonomy and resilience of the Euro-pean payments system, both of which underpinthe euro’s international role (Cœuré 2019c).The second is the rapid technological trans-formation that has occurred in peoples’ lives.The third focuses on the public criticism to therole of central banks due to their handling ofthe recent financial crisis (Cœuré 2019b). Thiscomes in contrast to the fact that they stand asa pillar of stability of the financial system andensure the smooth and secure functioning ofthe European payments markets. Thus, byweighing the costs and benefits, central bankshave to reassess and redefine their role in thepayments ecosystem (Habermeier et al. 2018;Cœuré 2019b).
These issues are explored in more detail in therest of the paper, which examines the prospectsof a retail account-based CBDC and is dividedinto four sections. Section 2 investigates thetaxonomy of money, the digital currencyecosystem and the main properties that a retailaccount-based CBDC has to fulfil. Section 3describes the main motivations for issuing anaccount-based CBDC as well as the existingcentral bank policies, and focuses on the pos-sible implications of such policies. In the lightof this analysis, Section 4 simulates theissuance of an account-based CBDC in the cur-rent Eurosystem’s payments environment.Finally, Section 5 provides some concludingremarks.
2 CBDC AND THE TAXONOMY OF MONEY
What will the money of the future look like?Today, one pays for a croissant at the cornerbakery by using either physical cash (ban-knotes and/or coins) or an electronic means ofpayment at the point of interaction (POI).Regardless of the price of the good or the cur-rency used to pay for it, an electronic paymentis usually performed by initiating a credit
transfer or a card payment. In the former case,one initiates the transaction by giving an orderto irrevocably debit one’s payment accountwhereas, in the latter, one uses a payment cardwith a debit, credit or delayed debit functionat the POI. However, one could pay for thegood just by waving a mobile phone or a wear-able to the terminal using near-field commu-nication (NFC) or augmented reality (AR)technology, scanning a quick response (QR)code or receiving a request to pay (RtP) andaccepting the payment. What ten years agowas considered science fiction is today’s real-ity. The digitalisation of payments has coin-cided with the dematerialisation of financialassets. Nowadays, the digital penetration inthe retail payments arena is such that, forexample, payments are regularly initiatedthrough mobile apps.
Indeed, digitalisation in the payments ecosys-tem has been widespread and profound. In thefront-end retail sector, numerous and fancysolutions are available to the customer whowishes to initiate a payment. At the back-end,Distributed Ledger Technologies (DLTs), thebest known of which is blockchain, promisegreater resilience, immutability, transactionprivacy, as well as higher speed and scalability(Gandal and Halaburda 2014; Robleh et al.2014; Middleton and Sinha 2019; Fleuret andLyons 2020). Nevertheless, DLT, althoughhighly promising, is not mature enough toreplace the traditional account-based RealTime Gross Settlement System (RTGS)3
(Kahn and Roberds 2009), which forms thecentral bank’s basic tool in order to implementmonetary policy and ensure financial stability(Kokkola 2010). In addition, DLT, althoughresilient to malicious incidents and able to con-front single point of failure issues, is currentlyusing consensus protocols4. Consequently, it
51Economic BulletinJuly 202094
3 In Europe TARGET2, the RTGS owned and operated by theEurosystem, is the system that settles payments related to theEurosystem’s transactions, as well as to bank-to-bank andcommercial transactions.
4 The author’s proposal, presented at the 1st EUROchainHackathon, for issuing retail account-based CBDC by usingpermissioned DLT, uses central bank blind signature to verify thetransactions in order to prevent double spending.
does not fully comply with the legal and regu-latory notion of ultimate finality, as requiredin the payment chain of a credit transfer, andalso suffers from interoperability and account-ability issues among the different DLT net-works5 (Tobias and Mancini-Griffoli 2019;Middleton and Sinha 2019; Ward andRochemont 2019). Irrespective of this argu-mentation and regardless of the technologyused at the back-end of a payment system, thequestion remains: What will the money of thefuture be like?
A number of fiat, digital currencies and pay-ment solutions, including EUR, USD, JPY,Bitcoin, Libra, Alipay, and Apple Pay, competeto dominate our physical or digital wallet6
(Tobias and Mancini-Griffoli 2019). Normally,questions are raised about the autonomy, trust,and habits of our everyday payment transac-tions. However, surprising as it may sound,what is money? Money is fundamental to theoperation of a modern economy and, strictlyspeaking, must serve three basic and not mutu-ally exclusive functions as: a unit of account, amedium of exchange, and a store of value(Quinn and Roberds 2014; McLeay et al. 2014;Bjerg et al. 2018). A unit of account refers toa standard numerical monetary unit of meas-urement of price or cost of goods and services,for example EUR 100 to buy and EUR 10 torent a bicycle, respectively. In order to suc-cessfully fulfil the aforementioned attributes,money must exhibit a stable price7. TheEurosystem, by targeting inflation, enhancestrust between the various economic agents andhelps them take rational economic decisions.As a medium of exchange, money is universallyaccepted as payment for goods and services. Asa successor of the barter system, it provides themost efficient, secure and universally availablemeans of exchange. Finally, as a store of value,money has to be able to be securely stored and,when subsequently retrieved, to retain a stablelevel of purchasing power. Moreover, apartfrom the codification of money, some simplefinancial concepts should be presented prior tothe analysis of the retail account-based CBDC.The reason is that “sound conclusions are
sometimes supported by erroneous arguments,and the error is compounded when a soundconclusion is declared to be mistaken on theground that the argument for it is mistaken”(Nagel 1963, p. 211).
Currency8 is in fact money in circulationdenominated in an existing unit of account, i.e.euro banknotes and coins, which serves as amedium of exchange (Kiyotaki and Wright1989; Yang 2007; Brunnermeier and James2019; Engert and Fung 2020). Thus, currencydoes fulfil the three basic functions of money.A further examination distinguishes betweentwo main types of money according to who theissuer is, i.e. central banks (monetary base) orcommercial banks (bank deposits). Cash isissued by central banks and remains the mostwidespread means of retail payment (Cruijsenvan der et al. 2015; Bech et al. 2018) in theeuro area9 (Esselink and Hernández 2017).The main reason for holding and using cash inour everyday transactions is trust. People trustthe central bank’s promise that a banknoteretains its face value, i.e. that it would beaccepted by others at its face value10, for exam-ple that a EUR 5 banknote buys EUR 5 worthof goods (Allen and Dent 2010). A second rea-son for using cash as a means of payment isbecause it constitutes fiat money by havinglegal tender status11. Fiat money has the uniqueproperty of being issued and accepted by ahigher authority (the government) as a meansof payment for salaries and taxes.12
51Economic Bulletin
July 2020 95
5 DLT is currently falling short in scalability, energy efficiency, andpayment finality (He et al. 2017).
6 For narrow finance solutions, see Pennacchi (2012) and Mancini-Griffoli et al. (2018).
7 Price stability is actually the primary objective of the ECB, whichis clearly established in the Treaty on the Functioning of theEuropean Union (2012, Article 2).
8 Deriving from the Latin word “currens” meaning: “in circulation”.9 Esselink and Hernández (2017, p. 19) argue that across the euro
area in 2016, on average, 78.8% of all POI transactions were carriedout using cash.
10 Greek drachmas bore the inscription “ΠΛΗΡΩΤΕΕΣ ΜΕ ΤΗΝΕΜΦΑΝΙΣΗ” meaning: “to be paid on demand”.
11 On 1 January 2002, the euro became the sole legal tenderthroughout the euro area, as the then President of the ECB, Dr.Willem F. Duisenberg, declared.
12 Usually, people erroneously believe that paying by cash at a shop,the merchant is obliged to accept it because it is legal tender. Thisis not true. Only central banks do accept cash. Legal tender meansthat if we pay a debt with euro banknotes, then the creditor cannotsue us for failing to repay.
What does actually make cash such a popularmeans of payment in everyday transactions? Inany case, taxonomy and terminology are unre-lated to habits and daily economic behaviour.Cash is the most popular instrument for retailpayments because it offers a real attractiveadvantage to users (Klee 2008; Arango et al.2015; Wakamori and Welte 2017). A cash trans-action offers full anonymity which, combinedwith convenience, is indeed a crucial benefitwhen using cash. Furthermore, cash has a num-ber of additional benefits. It offers the capa-bility to transact in real time, 24/7/365, irre-spective of payment systems, intermediaries andinfrastructures. In addition, it provides imme-diate and final settlement, with no single pointof failure, and no settlement and cyber risk asboth transacting parties are physically present(Weber 2015; Grym et al. 2017; Masciandaro2018; Middleton and Lund 2018; Middleton andSinha 2019; Tobias and Mancini-Griffoli 2019).Finally, it is possible for a person to negotiatea lower price when paying for goods and serv-ices in cash, also avoiding possible extra creditcard surcharges (Rogoff 2014; Wakamori andWelte 2017).
However, cash can be used mainly for prox-imity payments, resulting in additional costswhen a person wishes e.g. to send money to rel-atives and friends who live far away. Moreover,its use is limited to low value transactions(Wakamori and Welte 2017), it is expensive tostore and handle, as well as heavy and bulky inlarge amounts, and costly to count and process.Additionally, it is vulnerable to theft and sus-ceptible to counterfeit (Middleton and Sinha2019). In financial terms, cash holders do notenjoy interest or dividend returns, as storedcash cannot be used for investment purposes.Regarding some current trends, cash is facingdecreasing acceptance, mainly in Scandinaviancountries, as well as restrictions on the maxi-mum value of cash transactions set by manygovernments.13 Finally, cash has a reputationalcost, as it is widely used in the shadow economyand remains a facilitating medium of criminalactivities such as money laundering, terroristfinancing and the financing of the proliferation
of weapons of mass destruction (Rogoff 2014;Wright et al. 2017).
It is important to understand that cash is a phys-ical means of payment that does not require theexistence of payment accounts to initiate a pay-ment. Additionally, banknotes in circulation arepart of central banks’ balance sheet liabilities.Thus, if one wants to pay in central bank money,one has to pay by using cash. In other words,including central bank trust in payments meansusing cash stored in wallets or withdrawn froman ATM (thus reducing a commercial bank’saccount balance). In any case, the majority ofmoney in the economy appears on the liabilityside of commercial banks’ balance sheets, sinceit is kept in commercial banks’ deposits whichpotentially can be converted into legal tenderand hence central bank liability (Kokkola 2010).However, commercial banks do settle inter-bank, customer, and securities14 related pay-ments, as well as monetary policy operations incentral bank money, i.e. money in an accountheld with a central bank15.
Currently, commercial banks’ deposits form themajority of monetary circulation in developedeconomies and can be safely kept in accounts,converted into legal tender, or used in varioustransactions (Bossone 2001; Kokkola 2010).The latter include credit transfers, instant credittransfers, card payments, and direct debits, i.e.the main instruments used in the euro area toinitiate electronic payments (e-payments)from payment accounts. Irrespective of the userpenetration and growth of said instruments, atthe local or the European level16, it should benoted that an e-payment transaction is not
51Economic BulletinJuly 202096
13 In Greece the maximum amount of using cash in paymenttransactions is restricted to EUR 500. Usually, such restrictions areplaced as measures against tax fraud and AML since the featureof anonymity encourages persons to conduct transactions in cash.
14 In Europe, securities transactions are conducted in central bankmoney. TARGET2-Securities (T2S) settles the transaction on adelivery-versus-payment (DvP) basis, i.e. money and securitieschange hands simultaneously.
15 The reserves are actually sight deposits held with the nationalcentral banks (NCBs) by credit institutions in order to settlepayments, and meet the minimum reserve requirements and thedeposit facility. Regarding allocating reserves to the general public,see Brunnermeier and Niepelt (2019) and Niepelt (2019).
16 Payment cards have overtaken credit transfers as the most widelyused cashless payment instrument in the euro area.
anonymous but traceable. However, the crucialbenefit of an e-payment, besides the obvioususer convenience it offers, consists in its scala-bility since it is used both for retail and largervalue payments. An e-payment serves not onlyproximity payments at the POI, like cash, butalso e-commerce transactions. Additionally, ane-payment also offers 24/7/365 transacting capa-bility and, in the case of the instant credit trans-fer, real time availability of funds to the payee17
(Mancini-Griffoli et al. 2018).
One may argue that the initiation of an e-pay-ment simultaneously creates the need for abacking payment system able to manage allforms of “settlement risk”, i.e. systemic, credit,liquidity, operational and/or legal risk(Kokkola 2010). It is common knowledge thate-payments are vulnerable to fraud, opera-tional failure, and cyberattacks. Yet it is impor-tant to insist that security, resilience and set-tlement in central bank money are of para-mount importance when initiating an e-pay-ment in the euro area (Cœuré 2019c; Bank ofEngland 2020). In this context, it is importantto stress that in Europe there is a single systemof banking supervision, ensuring the safety andsoundness of the European commercial bank-ing sector, a single framework for resolvingfailing commercial banks, with minimal costsfor taxpayers and the real economy, and a sin-gle resolution fund financed by the commercialbanking sector18 (Cœuré 2019c).
E-money19 payments are currently of marginalimportance and magnitude in the euro area(Tobias 2019). The ECB defines e-money as “anelectronic store of monetary value on a techni-cal device that may be widely used for makingpayments to entities other than the e-moneyissuer” (European Central Bank 2020a). Con-sequently, e-money transactions do not neces-sarily involve commercial bank accounts.Instead, they involve prepaid instrumentsmostly used to execute small value retail trans-actions (Chiu and Wong 2014, 2015). The userinitiates an e-money payment from a digital wal-let which stores: (i) secure information neces-sary for authenticating the user; and (ii) pay-
ment data necessary for initiating a paymenttransaction usually through a mobile-basedapplication. The e-money directive (EMD) reg-ulates the storage and use of e-money in retailpayments. The result is a payment experiencevery similar to the use of a traditional com-mercial bank account. The key differencerelates to the nature of the transaction, as e-money is neither a unit of account nor a wide-spread store of value, and by definition nor legaltender. Usually, full backing in local fiat cur-rency, through deposits held with a supervisedcommercial bank, enables e-money institutionsto retain one-to-one parity with said currency,i.e. the euro in the case of the euro area.
A cryptocurrency20, such as Bitcoin, Ethereumor Ripple, has its own denomination, is cre-ated/mined by non-trusted intermediaries andis issued on a blockchain21 (Milne 2018; Pich-ler and Summer 2018; Fatás 2019). In thisrespect, it is important to note that cryp-tocurrencies (unlike commercial bankdeposits) are not easily converted into fiat cur-rency and most importantly they are nobody’sliability. In fact, even when they perform someof the functions of money, they remain specu-lative assets with unstable value that sufferfrom excessive volatility22 (Yermack 2013).
51Economic Bulletin
July 2020 97
17 Visa by launching a new service, Visa Direct, intends to offer a cardpayment solution very similar to an instant payment.
18 Single Supervisory Mechanism (SSM – Regulation (EU) No468/2014), Single Resolution Mechanism (SRM – Regulation (EU)No 806/2014).
19 The e-money directive (EMD) sets out the rules for the businesspractices and supervision of e-money institutions.
20 As mentioned above, we rather consider cryptocurrencies ascryptoassets since they do not satisfy the three basic functions ofmoney.
21 Blockchain refers to all types of distributed ledger technologies(DLT) including the protocol, network, and application layers.DLT refers to a combination of technologies that together createa digital, shared and automatically updated ledger of verifiedtransactions or information among parties in a network. DLTs aredesigned to operate with varying degrees of control, ranging fromcentralised models through to instances where there is no controlby a central authority(ies).
22 Currently some cryptoassets have been classified as “stablecoins”.For example, Facebook, driven by strong network effects, seeks tostabilise Libra’s price by linking its value to a basket of stablecurrencies. As noted in the Introduction, a detailed analysis ofstablecoins is beyond the scope of this paper. However, the authordoes agree with the outcome of the G7 Working Group onstablecoins that “their adoption is, as yet, uncertain as they facesignificant legal, regulatory, supervisory and operationalchallenges. Stablecoins, regardless of size, pose challenges and risksto AML/CFT efforts across jurisdictions, as well as operationalresilience (including for cyber security), consumer/investor and dataprotection, and tax compliance” (BIS 2019, p. 20).
Nevertheless, the main attractive feature ofcryptocurrencies is their ability to offeranonymity of transactions to their users (Ali etal. 2014a, 2014b; Auer 2019). However, theyare not used in large value transactions and areaccepted only by a limited number of retailers(Biais et al. 2019). In addition, they do not fullysatisfy the function of a medium of exchangesince limited acceptance prevents them frombeing a universally accepted means of payment(Mancini-Griffoli et al. 2018). Finally,anonymity of transactions often facilitatesmoney laundering and financing of terrorismand hence poses a direct threat to economicstability (Fung and Halaburda 2014; Bank forInternational Settlements 2015; EuropeanCentral Bank 2015).
CBDC is defined as the digital cash which canbe issued by central banks for payments andsettlements (Bjerg 2017; Adrian and Mancini-Griffoli 2019). Thus, if one wanted to use cashin digital form, then one would have to useCBDC (called here d-euro (DEUR) in the caseof the Eurosystem). The relevant literature(Bech and Garratt 2017; Mancini-Griffoli et al.2018; Barontini and Holden 2019; Carstens2019; Kahn et al. 2019) divides CBDC into twomain categories, according to the accessibilitycriterion, i.e. retail and wholesale CBDC.Retail CBDC would offer public access to cen-tral bank liabilities and can additionally bedivided into two sub-categories based on theverification of the transaction. The retailtoken-based CBDC is the case of central banksproviding the general public with direct accessto CBDC23 (World Economic Forum 2020). Inthis direct tokenised form, CBDC can bestored either on a mobile phone or on a pre-paid card. Retail account-based CBDC24 is cre-ated when commercial banks provide the pub-lic with indirect access to CBDC through a spe-cific type of commercial bank accounts25 (Fungand Halaburda 2016; Bjerg 2017). In this indi-rect access form, central banks would create,maintain, hold, regulate and operate theCBDC accounts for commercial banks whilethe latter would provide the client interface,such as know your customer (KYC) and front-
end innovative payment solutions26. This typeof CBDC involves a partnership between pub-lic and private sectors (Lagarde 2018) whoseeffectiveness is enhanced by their separate anddistinct roles (Kashyap et al. 2002).
In another type called wholesale token-based27
CBDC, central banks provide commercialbanks and/or other financial institutions withrestricted direct access to accounts where theycan settle interbank payments (Bank for Inter-national Settlements 2018). Such an innovationcould improve cost efficiency, increase speed,and provide 24/7 access to payments. However,discussing the merits and costs of wholesaleCBDC as well as its monetary policy implica-tions lies beyond the scope of this paper. Thelatter focuses only on the implications of acommercial bank account-based CBDC pro-viding access to central bank money to the gen-eral public, instead of a selected number ofusers (e.g. eligible counterparties that have theobligation to hold minimum reserves).
The following figures summarise the foregoingdiscussion. Figure 1 presents d-euro within thetaxonomy of currently existing forms of money,on the basis of access criteria. Figure 2 showsan altered version of the money flower origi-nally designed by Bech and Garratt (2017).This version highlights the property of con-vertibility, as shown in the red petal. In fact,the coexistence of central and commercialbank money requires that they are convertibleat par, leaving the quantity of central bankmoney unaffected. Figure 2 stresses the fact
51Economic BulletinJuly 202098
23 According to Kahn and Roberts (2009, p. 11), the main differencebetween an account and a token lies in the verification needed,since the account holder is verified through a messaging processand the token’s validity is verified by the beneficiary (Bank forInternational Settlements 2018, p. 10).
24 Sometimes referred to as a “synthetic CBDC” (Tobias 2019).25 An alternative (not supported by the author) is a retail direct
account-based CBDC when central banks offer to the generalpublic direct access to accounts.
26 The author’s business model of issuing a retail account-basedCBDC, presented at the 1st EUROchain Hackathon, proposed aspecial type of account called “Dedicated Digital Account” (DDA)for d-euro transactions via the TARGET TIPS service (TARGETInstant Payment System). A detailed outline is provided in Section4.
27 Double spending is the main worry in the digital world, i.e. whetherthe token has already been spent.
that cryptocurrencies, including Bitcoin, areinconvertible and not backed by any asset,unlike commercial bank money.
A close observation of the proposed d-euro asa retail account-based CBDC shows that itsatisfies three of the four properties presentedin the money flower diagram: it is convertible,it is digital and it constitutes a central bankliability.
For an account-based d-euro to be availableto the general public, six principles have to be
fulfilled (Bank for International Settlements2018).
• Principle 1: Anonymity. This is the featurethat makes cash the most attractive and pop-ular means of payment nowadays (Athey etal. 2017). Accordingly, d-euro must offer acompetitive degree of anonymity as a vitalattractive feature to potential users(Mancini-Griffoli et al. 2018). In contrast tocash, which is fully anonymous and untrace-able, providing an advantage in terms of privacy, the use of cash-like d-euro must
51Economic Bulletin
July 2020 99
€
Figure 1 The proposed d euro wit
€
€
hin the taxonomy of currently existi
ng forms of money
setoknnBa setoknnBad-euro
ermuminiM(whod-euro
sgnivaSsevreseolesale)
sgnivaSdepositsTime
e-money errucotpyCry seicnerrucotpyCr
Figure 1 The proposed d-euro wit
hin the taxonomy of currently existi
ng forms of money
Figure 2 d-euro and the money flo
brevnoC
tigiD
le
d d
ower
tal e
koTo
ytilibaliknablartneC
desab-nekoTTo
)na )desab-tnl
uocca(r iatre
o- rue-
s
d
desae
b-nekotalaso
eolhw(orue-d
hsaC
)desab-nekotliat(reorue-d
)
be semi-anonymous (Lagarde 2018)28
(anonymity may be lifted due to AML con-cerns29 – Garratt and van Oordt 2019; Euro-pean Central Bank 201b, 2019c). This semi-anonymity30 property makes d-euro an idealalternative anti-money laundering solutionunlike physical cash.
• Principle 2: Trust. The account-based d-euromust be fiat money with legal tender status.As by definition d-euro would be a liabilityon a central bank’s balance sheet, it shouldinspire the same level of trust as cash. Suchtrust would emanate from the fact that d-euro would be issued by the central bankand would be convertible at par with allother forms of the euro31. The Eurosystem’snational central banks (NCBs) would carryout the task of persuading the general pub-lic that d-euro’s value is the same as that ofeuro banknotes and coins. It is envisagedthat, by achieving a stable and safe asset sta-tus, the d-euro will create positive reputa-tion and trust spillover effects for centralbanks. The latter are seen by the public “aspart of a financial system which has failedto deliver growth and fairness” (Cœuré2019b). Central banks should view theissuance of d-euro as an opportunity torespond to the rapid evolution of digitalreality, the decline in the use of cash, as wellas the multiplicity of FinTech solutions (Heet al. 2017; Cœuré 2019b; Fatás 2019) andprivately issued cryptocurrencies. Centralbanks must not be left behind but insteadexploit their incumbent position and issuedigital trust themselves. As Europeans showa positive attitude towards holding paymentaccounts, there is a significant potential fora public-private partnership in this area(Lagarde 2018; Bank of England 2020). It isthe only way for central banks to underpintrust32 in fiat currency and disincentivise thegeneral public from adopting private sectorfront-end solutions33.
• Principle 3: Scalability. The account-based d-euro would be used mainly for person-to-person (P2P), person-to-business (P2B), and
person-to-government (P2G) payments(like physical cash). The pivotal role of d-euro lies in its efficiency and scalability inboth online and POI payments. Theissuance of the d-euro could play a wide-spread role in the use of the single currencyin the European retail payments market,thereby contributing to the autonomy of theEuropean payment systems as well as pro-moting the international role of the euro. Inother words, d-euro should form the cor-nerstone both of the digital approach to theeuro and of the effort to enhance its positionas an international currency34. Therefore, itis expected to be a decisive step in theEurosystem’s attempts to meet the increas-ing needs of consumers for efficient andsecure domestic and cross-border retail pay-ments complementing already existingsolutions35 (Hasan et al. 2013; Cœuré2019c). Summarising, scalability, combinedwith trust, will most probably be the decisivefactors behind the wide adoption of d-euro.
51Economic BulletinJuly 2020100
28 The author’s business proposal maintains semi-anonymity, as theNCB which participates in the consensus-driven validation processdoes not see the content of the transaction. The latter can be seenonly by order of the AML officer.
29 In principle, under Article 2(2) of the 4th EU AML Directive (EU2015/849), cash transactions are out of scope when there is not anyintermediary intervention.
30 It should be underlined, strictly speaking, that in a given d-europayment transaction the identity of the payer and the payee shouldbe fully anonymous to all parties involved. Only the central bankcan retrieve such information if and only if there is an AML officerdecision, and given that there is reasonable suspicion of an illegaleconomic activity. In addition, at the 2nd EUROchain Hackathonin October 2018, the author’s proposal was that the account-basedCBDC should be fully anonymous for transactions of up to DEUR500 (currently in Greece any P2B transaction in excess of EUR 500is only allowed via e-payment) and semi-anonymous for paymentsabove that ceiling.
31 The author’s business proposal outlined the necessity of pricestability in all economic circumstances and hence of the one-to-oneparity rule vis-à-vis the physical euro.
32 Because the value of money resides in trust, one may argue thatonce people start using private money, they may lose a degree oftrust in the sovereign currency.
33 Mobile payments platforms that are based on the network effect,like Apple Pay, have been witnessing a significant growth in thepayments landscape.
34 Given that the US dollar is the dominant international currency,private initiatives have a competitive advantage in front-endsolutions, brand loyalty and global acceptance, compared withcentral banks and as stated by Gresham’s law, “bad money drivesout good money” (Giffen 1891, p. 304), the issuance of the d-euroshould be seen as the right step to a “euroized” digital economy.
35 Cœuré (2019a) has argued that “it is probably easier to connect anew currency to an existing network – the case of Libra – than tobuild a new network on an existing currency – the case of the euro”.In other words, the issuing of [the d-euro] achieves a hybridsolution: it introduces a new digital currency to an existing network,i.e. TARGET services.
• Principle 4: Interest rate. The account-basedd-euro should be interest rate free36 andhence unable to function as a monetary pol-icy tool37. Such a semi-anonymous d-euro,issued by the central bank, with a zero nom-inal return embedded in its initial design, ismore trustworthy and likely to be adoptedthan most private cryptocurrencies. Thus,under normal economic conditions, d-euroshould be designed so as to set the effectiveinterest rate lower bound equal to zero (Hall2013; Agarwal and Kimball 2015; McAn-drews 2015; Dyson and Hodgson 2016;Engert and Fung 2017; Meaning et al. 2017;Middleton and Sinha 2019). However, thatbeing said, the decision of whether or notthe d-euro will be interest-bearing is an issuewith multiple implications for monetary pol-icy and financial stability (Eggertsson andWoodford 2003; Mancini-Griffoli et al.2018; Agur et al. 2019; Keister and Sanches2019; Kim and Kwon 2019).
• Principle 5: Acceptance. Initially, theaccount-based d-euro should be used in pay-ment systems and by payment solutions thatoffer pan-European reachability with globalacceptance as a long-term goal38 (Cœuré2019c). Global acceptance should reinforceeconomies of scale, support the euro’s inter-national role, and meet the increasing needsof European consumers for global reacha-bility. In fact, Europe has made substantialprogress in the promotion of the euro as aninternational currency since its introductionon January 1, 2002. Cash payments havebeen carried out using the single currencyacross the euro area with the same ease andconvenience, but this has not been the casefor electronic payments. The developmentand implementation of the Single Euro Pay-ments Area (SEPA) in 201439 resulted in thecreation of a single payments area for cash-less payment instruments. Since the creationof SEPA, the d-euro seems the natural nextstep towards pan-European acceptance andreachability for both proximity and e-com-merce payment transactions. While GAFA40
are leveraging their extensive networks and
are increasing their presence in the front-end retail payment environment in Europe,the issuance of the d-euro could be seen asturning this challenge into an opportunity.
• Principle 6: Convenience. The account-basedd-euro should be designed according to cus-tomers’ demands for convenience, speed,efficiency and global acceptance for bothproximity and e-commerce payments. The d-euro issuance should be seen as a possibleresponse of the Eurosystem to private sectorstablecoin and cryptocurrency initiatives andcard scheme global players (Cœuré 2019c).The d-euro has to cater for consumers’changing needs by combining the character-istics of cash and payment cards41. Thus, itmust be user-friendly, equal to availableinstant payment solutions, and capable tooffer 24/7/365 payments in real time. In addi-tion, it should be available through differenttools, payment instruments, and technolo-gies, enhancing the payment experience andbecoming synonym of convenience to bothconsumers and merchants. So far we haveargued in favour of a public-private part-nership between central and commercialbanks, capable of creating a pan-Europeanretail payment solution, like a European-gov-erned payment card scheme, where centralbanks provide the back-end infrastructureand the private sector the front-end solutions.
As already stated, a CBDC initiative will haveto serve the three basic functions of money, i.e.
51Economic Bulletin
July 2020 101
36 The author’s business proposal for the 1st EUROchain Hackathonoutlined the necessity of the one-to-one parity rule vis-à-vis thephysical euro and hence argued for a non-yield-bearing CBDC.
37 The next section considers the possibility of using the d-euro as amonetary policy instrument.
38 The key advantage of the stablecoin initiatives, based on theirglobal clientele, is their cross-border functionality.
39 SEPA is an area where consumers, enterprises and public entitiescan make and receive domestic and cross-border electronicpayments in euro under the same basic conditions and with thesame rights and obligations. SEPA has created an integrated,competitive and innovative retail payments market, where allcashless euro payments are conducted via a single payment accountand a single set of payment instruments in a fast, safe and efficientmanner.
40 The non-European tech giants Google, Apple, Facebook andAmazon are often referred to collectively as “GAFA”.
41 A payment card with a common European brand and logo, openaccess to NFC interfaces for proximity payments, and a globalacceptance for e-commerce should be an efficient solution.
stable unit of account, costless medium ofexchange, and secure store of value (Bordo andLevin 2018). Additionally, and in view of theaforementioned set of principles, the case forthe d-euro raises a number of new policy con-siderations that merit further investigation(Andolfatto 2018).
3 CBDC AND THE ROLE OF CENTRAL BANKS
It is worth noting that digitalisation in the fieldof payments has affected consumers’ paymentbehaviour (Cœuré 2019a). However, highdemand for digital payment solutions mayquestion public trust in major financial insti-tutions. Consequently, many central bankshave realised that they should respond imme-diately to the recent technological innovationsin banking and payment systems, as well as tothe digitalisation of money (Ahmat and Bashir2017; Engert and Fung 2017; Bergara andPonce 2018; Juks 2018; Kumhof and Noone2018; Panetta 2018; Sveriges Riksbank 2018b;Norges Bank 2019; Bank of England 2020). InEurope, the Eurosystem has started to con-sider the issuance of a CBDC and to analyse itspossible design and implications (Cœuré2018). Thus, digitalisation has challenged therole of central banks, to the extent that theyhave to be involved in the retail payments mar-ket and to act beyond their traditional role asguarantors of price and financial stability andensurers of the smooth operation of the pay-ment systems. Central banks are presentedwith the significant opportunity of analysingthe potential implications and risk considera-tions arising from the issuance of a digital cur-rency such as the d-euro. By doing so, and bydeciding that they must not be left behind inthe digital race, they are also faced with a keyquestion: what are the main motivationsbehind the issuance of an account-based retailCBDC and why might central banks wish toproceed with it?
Motive 1: Trust and the digital transformationof the economy. Facebook’s Libra, based onpermissioned blockchain technology (Libra
2019), is a token-based stablecoin backed by a pool of assets. Ιt seeks to reach wide-scaleadoption and global acceptance and expects tobenefit from network effects and economies of scale (Cœuré 2019a). Indeed, Facebook(which claims 2.45 billion core active users in201942, as well as 2 billion WhatsApp and 1 bil-lion Instagram users) tailors its payment solu-tions to consumers’ changing behaviouraltrends and attempts to turn a means of pay-ment into a global currency (Gans and Hal-aburda 2013; Fung and Halaburda 2014;Cœuré 2019a). By entering the realm of retailpayments and financial services, Facebookplans to provide both front-end and back-endsolutions, thus eliminating the need for centralclearing and settlement performed by centralbanks. Furthermore, Alipay and WeChatPay,43 using mobile payment solutions andnon-social networking apps, together accountfor 93% of all digital transactions in China.44
These third-party online and mobile paymentsolutions strive to become globally available bysigning deals with European banks and digitalwallet companies. By doing so, they intend toprovide tailor-made services meeting theneeds of European customers.
Moreover, Google Pay and Apple Pay arecard-based wallet solutions that allow users toinitiate proximity payment transactions at thePOI and are prominently focusing on e-com-merce payments. Thus, they offer services thathave been traditionally provided by the bank-ing sector. Given their branding power, theyare gradually replacing the existing front-endbanking solutions and provide additionalback-end core banking services which usuallylie outside of the banking perimeter (Cœuré2019b). In addition, as these solutions are fast,cheap, user-friendly and offer cross-borderfunctionality, they may eagerly be adopted byEuropean customers for purchases at domes-tic or foreign merchants.
51Economic BulletinJuly 2020102
42 The world’s population at that year was 7.71 billion.43 Alipay and WeChat Pay provide mainly front-end services, while
at the same time decreasing merchants’ costs by using the QR codetechnology, which is cheaper than the card-based technology.
44 According to the People’s Bank of China annual report.
Alipay, WeChat Pay, Google Pay, Apple Pay,and many others are targeting to shift con-sumers’ payment preferences by importing pro-prietary practices that are alien to Europeanconsumers’ habits and needs. In an extremescenario, given information asymmetry andadverse selection, based on a winner-takes-alleffect, the Eurosystem’s market infrastructuresservices, TARGET2 and TIPS, may beimpacted. In this case, the bigtech firms thatoffer payment solutions may exploit their com-parative advantages, bringing about a signifi-cant decrease in the share of euro area banksin the European retail payments market andaccordingly a tantamount drop in the volumeof assets settled in central bank money, viaTARGET2. Given that under the Europeanpayments strategy, the pan-European solutionmust be market-led and meet the objectives ofthe Eurosystem, the foregoing solutionsoffered could cause an adverse impact on stan-dardisation, interoperability and reachability inthe European retail payments market.
By acting as operators, overseers and catalysts,central banks are responsible for providingefficient and sound clearing and payments sys-tems as well as for maintaining the value andtrust of the currency (Kokkola 2010). The safeand efficient use of the currency as a stablemedium of exchange is a basic pillar of con-sumers’ trust in it. The fact that a privatelyissued stablecoin may replace traditional pay-ment and financial assets is a wake-up call forcentral banks to start considering issuing a dig-ital alternative that warrants maximum reach-ability, stability, and resilience. The d-euroshould be an ideal answer, as it enables pan-European SCT-Inst-based45 POI solutions viaTIPS. By offering such pan-European reacha-bility, it minimises any potential reputationalrisks regarding the role of the Eurosystem inthe payments ecosystem.
In order to be prepared for the advent of 5Gtechnology and to start promoting climate riskdisclosure in their tasks, central banks have tobe ready to take the leadership regarding theprovision of technological innovation. In this
context, the potential issuing of the d-euro canbe seen as the next step in the digitalisationprocess of the retail payments market. Centralbanks would thus become active players in theprocess of money digitalisation in order to fos-ter trust and to promote a reliable and efficientmeans of payment. In other words, the issuanceof the d-euro as the digital version of a sover-eign currency, i.e. the euro, would support theresilience and strength of the European mon-etary policy project. In addition, it would sup-port the safety and soundness of the Europeanretail payments market and could promote theinternational use of the euro as a stable andconvenient currency. It is a well-establishedprinciple that “it is difficult to build a new net-work on an existing currency”46. The Eurosys-tem has created TIPS, a TARGET service, withthe potential to ensure full pan-Europeanreachability and settlement of instant paymentsin d-euro as well. Thus, and since TIPS has amulticurrency technical capacity, d-euro can beintroduced from day one as a stable means ofpayment with immediate pan-European reachand potential global acceptance.
Motive 2: Decline in the use of cash. Cash is thedominant payment instrument in the euroarea. However, an extreme scenario couldmaterialise in the future where physical cashwill not be generally accepted as a means ofpayment and demand for it will decrease. Cur-rently, in some countries, most notably Swe-den, the use of physical cash for retail paymenttransactions has been declining rapidly and anumber of retailers do not accept cash at all47
(see for example Bordo and Levin 2017). Addi-tionally, according to the latest payment sta-tistics,48 the euro area has experienced a steadydecline in the number of bank-operated ATMsand an increase in the number of card pay-
51Economic Bulletin
July 2020 103
45 SCT Inst stands for SEPA Instant Credit Transfer.46 See footnote 36.47 The discussion of a cashless society goes beyond the scope of this
paper (for a discussion about this subject-matter, see Engert et al.(2018a) and Gnan and Masciandaro (2018)). The Eurosystem treatsall available means of payments on a non-discriminatory basis. Theauthor does not entertain the idea that cash will be abolished andhence any theoretical discussion about cashless societies is atpresent meaningless.
48 ECB Payment Statistics.
ments. These trends are likely to further accel-erate in the future, given the widespread adop-tion of open banking solutions and the intro-duction of new services offered by regulatedthird party providers (TPPs)49, especially thepayment initiation service providers (PISPs)50.The declining use of physical cash observed inmany northern European countries may be awarning signal about possible structuralchanges in consumers’ behaviour at the euroarea level. Given the proactive stance that cen-tral banks are likely to adopt in the new digi-tal paradigm and as Europe is moving towardsa cashless payment environment (Engert et al.2018a), one may ask: how will they fulfil theirmandates of providing resilient, efficient andstable payment systems?
Cash constitutes a fraction of a central bank’sbalance sheet and of the money supply (M1).Its popularity rests on the fact that it is risk-free central bank money that settles paymentswith anonymity and finality. In addition, strictlymicroeconomically speaking, a cash payment51
is a zero-sum game, where both parties (payerand payee) enjoy Pareto optimality, since abanknote is exchanged in order to redeem anobligation. Cash plays no role in the operationof TARGET services. There, payment ordersare submitted to the platform for processingand are irrevocably settled via risk-free trans-fers between the respective dedicated cashaccounts that the participants hold with thecentral bank.52 Thus, a significant decline (orincrease for that matter) in the use of cash can-not affect the back-end payment operations ofthe Eurosystem. Any change in the volume ofphysical cash in circulation will neitherobstruct nor interrupt the continuous opera-tion of TARGET services and hence will notaffect the primary objective of central banks topromote the smooth operation of the paymentsystems and to ensure efficiency and soundclearing (Kokkola 2010). However, thedemand for small-denomination coins is ratherlow as they are not used very often (Judson2018). Hoarding of coins increases the oppor-tunity cost of using cash against all otherinstruments and holding cash for saving rea-
sons53 due to its physical nature.54 It can beargued that this decline in the use of physicalcash may break the bond of trust between thepublic and the central bank, as the former losesconfidence and starts to question the latter’smandate and role (Middleton and Sinha 2019).The trust that people place in cash is merely areflection of their trust in central bank issuers.Hence, the issuance of the d-euro, as digitalcash, could be seen as a credible way of avoid-ing reputational risks and regaining trust.
It must be noted that as part of a portfoliodiversification strategy, people may hold cer-tain amounts of cash as a precautionary reservesince it is the most trusted form of money. Infact, it has been observed that demand for cashincreases dramatically during bank runs55 asthe public becomes more risk averse and looksfor risk-free assets (Brown et al. 2017). Duringthe recent euro area crisis, two major eventswere observed in Greece which was also goingthrough a sovereign debt crisis: (i) an increasein demand for cash as a safe asset; and (ii) arealisation that access to commercial bankdeposits is subject to regulatory and liquidityrestrictions. In June 2015, capital controls aswell as regulations enforcing the use andacceptance of electronic payment instruments(mainly cards) were imposed in order to stemcapital flight. Argentina has experienced a phe-nomenal demand for foreign currency (i.e. theUS dollar) and safe assets during its last eco-nomic crisis, as they were perceived as morestable and secure. Recommendations for thepromotion of electronic payment instrumentsin a crisis environment are neither addressedonly to the technology-adept generations, norare they necessarily related to financial rea-sons, as the COVID-19 pandemic has revealed
51Economic BulletinJuly 2020104
49 The EU’s revised Payment Services Directive (PSD2).50 The PISPs are authorised to initiate payments from a user’s
payment account.51 In strictly economic terms, a payment is indeed a transfer of funds
by using either cash or deposits held with credit institutions anddischarges an obligation on the part of the payer vis-à-vis the payee.
52 T2S settles securities transactions on a delivery-versus-payment(DvP) basis using optimisation and auto-collateralisationmechanisms to enhance settlement efficiency.
53 Store of value.54 The cost of cash has been described in Section 2.55 Financial crises have very low probability tails (Engert et al. 2018a).
with the nationwide lockdowns, the restrictionof movement, and the “stay home” campaign.56
All the cases discussed above exhibit one com-mon characteristic: at the moment physicalcash has proved to be the only form of risk-freecentral bank asset available to the public. Inparticular, a decline in the use of cash gradu-ally prompts consumers to depend greatly onpayment solutions provided by the private sec-tor and to lose their trust in efficient and reli-able payment systems. Therefore, centralbanks, by issuing the d-euro as a semi-anony-mous digital means of payment alternative tocash, would provide the public with access tocentral bank money that is a secure store ofvalue and a stable means of payment in peri-ods of financial (Mancini-Griffoli and Ranaldo2011) and societal crisis. After all, is this notone of their roles?
Motive 3: Financial stability considerations.57
One of the main issues related to financial sta-bility concerns possible bank runs as a conse-quence of the d-euro’s competition with com-mercial banks’ deposits (Stein 2012; Dyson andHodgson 2016; Raskin and Yermack 2016;Bank for International Settlements 2018). Suchcases have attracted significant interest in therecent literature and policy debates and can besummarised in two main scenarios: (i) a surgein conversion of commercial bank deposits intod-euro during normal financial times; and (ii)conversion of commercial bank deposits into d-euro (rather than cash) during periods offinancial crisis. It must be noted that in thisanalysis, d-euro deposits are interest-free ascredit is provided entirely by commercialbanks. In both scenarios the d-euro accountsare held with and managed by commercialbanks, while deposits are guaranteed, up to alimit, by national deposit guarantee schemes58.
The first scenario presupposes informationasymmetry between account holders, whichleads a part of them to hold d-euro as a saferstore of value and hence to increase itsdemand. In such a case, central banks mayintervene by issuing d-euro up to a specific
amount per holder, providing a certain hori-zontal quota59. Thus, by imposing ceilings onthe amount of d-euro allowed per customer,central banks can promote the currency’s func-tion as a medium of exchange instead of a storeof value.60 However, one could argue that com-mercial banks will react to the introduction ofthe d-euro by adjusting deposit rates even iflimits are imposed on d-euro account holders,thereby leading to an insignificant deposit out-flow into d-euro under a financial system inwhich the same authority guaranties both typesof deposits. Commercial banks are expected toraise revenues from the management ofaccounts (in the form of fees, etc.) as well asfrom a number of customer-tailored paymentsolutions that will be provided and treated aspremium services. Commercial banks willprobably react and offer additional services toaccount holders, similar to those provided tocard holders in order to make deposits moreattractive to potential customers. Finally, as analternative measure, they may raise interestrates on saving deposits in order to win clien-tele from their competitors. Such increasesmay spark a hike in lending rates, in order toreadjust the interest rate margins to their ini-tial level. In an extreme scenario, commercialbanks should consider replacing part of theiroutflows of retail deposits with secured whole-sale funding in order to reduce their depend-ence on them.61 Central banks may consider
51Economic Bulletin
July 2020 105
56 The decline in the use of cash experienced during the COVID-19pandemic has helped to prevent the spread of the disease to acertain degree. Banknotes are coronavirus carriers and people optfor contactless payments instead (COVID-19 stays on a note’ssurface for a number of days and thus people using cash have towash their hands regularly). However, attitudes adopted duringsuch emergency periods may have a more permanent nature andcan thus change consumers’ payment habits, leading to a fasterreduction in cash use.
57 Future work could consider a deeper analysis of this issue.58 In the EU, deposit guarantee schemes protect depositors’ savings
up to EUR 100,000 and prevent the mass withdrawal of depositsin the event of bank failures.
59 Assuming there is parity between all forms of money, i.e. cash andcommercial bank deposits, quotas must aim to enhance theefficiency of d-euro as a means of payment.
60 The author’s business proposal at the 1st EUROchain Hackathonstressed that for social reasons this quota should be closelyassociated with the total amount of euro banknotes and coins incirculation. Thus, according to this proposal, the Eurosystemshould introduce a single tier system for the account-based d-eurowith a dynamic ceiling, given the decrease in the use of physicalcash. The proposal called for an initial issuance of DEUR 4,000 percitizen.
61 Commercial banks with high retail deposits are more sensitive.
continuing to provide liquidity while acceptingriskier collateral and thus a higher degree ofcredit risk. Theoretically speaking, this situa-tion could lead to higher costs of funding,reduced and riskier profits and hence bankfunding problems and financial instability, asd-euro deposits will not be available for lend-ing. Yet again, given the low d-euro ceiling peraccount holder, the likelihood of reduced bankprofits and undermined financial stability isarguably low in an environment where centralbanks and/or prudential authorities safeguardthe stability of the financial system and oper-ate in partnership with commercial banks.
So far we have argued that during normalfinancial times, deposits in central bankmoney, held with and managed by commercialbanks, are both risk-free and perfectly liquid.However, during periods of financial crisis, cit-izens may perceive the d-euro as a safe-haven,risk-free store of value backed by the centralbank, and when in panic they may cause a bankrun. The Greek sovereign debt crisis, which ledto a series of austerity measures and the debtswap through private sector involvement (PSI),has shown that risk-free assets do not exist ina financial crisis environment (see Jobst andStix 2017). During periods of financial crisisand panic, d-euro may be perceived as an“asset of the last resort” offered by the centralbank. Two major potential benefits from issu-ing d-euro become apparent in times of eco-nomic turbulence: (i) given that Greece asmember of the euro area has lost its monetaryflexibility, liquidity may stay longer in thecountry, as capital flight may, to some extent,be avoided (capital outflows would severelyincrease the NCB’s liability vis-à-vis theECB62); and (ii) trust in the central bank andthe euro as legal tender may be underpinned,63
thus averting any additional reputational costs.Before the Greek experience of capital con-trols, a shift towards safe assets (notably cash)had been observed. Central banks could also,in future, consider introducing a two-tier d-euro system in order to meet money demandfor precautionary reasons (Bindseil 2020).Thus, besides the d-euro used for payments,
central banks could introduce a tier-two d-eurowhich would act exclusively as a store of valueto households and enterprises. The amount ofeach tier allocated to each account holderwould be determined and subject to change bythe central bank. The central bank would alsobe able to change the interest rate (Lenza et al.2010) of tier-two deposits at will (Bindseil2020), applying even a negative rate duringnormal times so as to make them highly unat-tractive. Nonetheless, the aforementioned con-cerns expressed in the literature regarding thepossibility of digital bank runs should informthe exact CBDC design.
Finally, d-euro is expected to be preferred overcash as a medium of exchange, due to non-financial reasons such as the time-consumingprocess of visiting ATMs or the cost of hold-ing cash (including the risk of being lost orstolen during transportation). As evidencedduring the very recent COVID-19 pandemiccrisis, the d-euro should appear to be suitableto serve as a semi-anonymous, cheap, secure,and efficient means of payments used forremote transactions (including e-commerce).
Motive 4: Monetary policy considerations. Thepotential impact from the introduction of thed-euro is expected to be of limited relevancefor monetary policy (see for example Bordoand Levin 2017). As long as physical cash is notabolished in the near future (Niepelt 2018), d-euro is both highly unlikely and undesirable(on the part of policy makers) to bear interestand serve as a monetary policy instrument (e.g.Hamilton and Wu 2011; Edgar 2012; Good-friend 2016; Rogoff 2016; Dell’Ariccia et al.2017; Davoodalhosseini and Mohammad2018). Conventional monetary policy dictatesthat when short-term interest rates are at ornear the so-called zero lower bound (ZLB),
51Economic BulletinJuly 2020106
62 https://www.bankofgreece.gr/en/main-tasks/payment-systems-and-settlements/target-services/large-value-payment-system-target2/target2-account-balance.
63 In situations of panic, d-euro could also be affected and treated asa “lemon” because of adverse selection according to the well-knownAkerlof’s paper (1970) and thus d-euro holders should also beexpected to withdraw their assets, thereby increasing the elasticityof demand for all deposits.
central banks cannot reduce them further inorder to stimulate lending and demand whenthe availability of zero return physical cashserves as a stable floor for all interest rates(Dyson and Hodgson 2016; Borgonovo et al.2018; Strournaras 2018). It can be argued thatas d-euro is the digital equivalent of cash, andits supply is constrained by design, then it actsas an interest-free asset and a concrete floor toall interest rates in the wider economy. In anegative interest rate economy, people willprefer to hold a zero-interest bearing asset likecash, d-euro, a foreign currency (e.g. USD) oreven a stablecoin, which should set the effec-tive lower bound to zero (Engert and Fung2017; Bordo and Levin 2018). By designing d-euro as an attractive complement to cash andkeeping its supply constant (at a certain ratioto the amount of cash in circulation), thereshould be no effect on short-term interest ratesand on the alleviation of the ZLB constraint.Thus, in fact, the circulation of physical cash(rather than the issuance of d-euro) is the mainbinding constraint to the application of nega-tive interest rates during periods of deflationand limited aggregate demand64. On the otherhand, one might argue that a zero-interest rated-euro would free the Eurosystem from theZLB issue. Niepelt (2018) argues that it wouldrelax central bankers’ ability to conduct mon-etary policy based on interest rate manipula-tion in order to stimulate the economy, therebyallowing them to focus on the primary objec-tive of the Eurosystem’s monetary policy, i.e.price stability.
Since it is not feasible to reduce interest ratesbelow certain limits, there are alternative mon-etary policy tools that the d-euro could serve,with a view to boosting aggregate demand.Quantitative easing (QE) has proven to be oneof the most effective policy tools in the euroarea during the financial crisis. When centralbanks apply quantitative easing, they buy secu-rities (e.g. government bonds) in order to pushdown interest rates and provide liquidity to thebanking sector (McLeay et al. 2014; Meaninget al. 2018). The ECB has recently announceda new asset purchase programme65 in response
to the COVID-19 pandemic, ensuring “that allsectors of the economy can benefit from sup-portive financing conditions that enable themto absorb this shock … equally to families,firms, banks and governments” (EuropeanCentral Bank 2020b). The implementation ofsuch a programme could be further facilitatedif people were able to hold d-euro accounts. Inanother case, central banks could perhapsmore easily provide liquidity directly to house-holds and other designated account holders(e.g. firms, pension funds, etc.) in what isknown as “helicopter money” (Reichlin et al.2013; Dyson and Hodgson 2016; Ward andRochemont 2019). Unlike QE where centralbanks purchase assets (using printed money)by increasing their liabilities, with helicoptermoney central banks make an impact on theeconomy by issuing money and distributing itdirectly to households and enterprises (Dysonand Hodgson 2016). Especially in cases like theCOVID-19 pandemic crisis when firms closedown and strategic industries suffer financiallosses or face bailouts, d-euro could facilitatethe implementation of a helicopter moneyscheme, which intends to help society copewith staff redundancies and lockdowns.
All in all, like cash, a d-euro would play arather minor role in monetary policy imple-mentation (Broadbent 2016; Agur 2018;Engert et al. 2018b; Albertazzi et al. 2020).However, d-euro may be used as a direct chan-nel through which central banks can transmitmonetary policy to households and enterprises.Central banks have been unwilling to open andmanage public d-euro accounts. Therefore, thistask should be assigned to commercial banks,along with the customer onboarding process aswell as the provision of front-end paymentsolutions. The next section presents our casefor a possible d-euro issuance that is based onexactly such a public-private partnership.
51Economic Bulletin
July 2020 107
64 Altavilla et al. (2019) argued that when the ZLB has been hit,healthy banks are better off, compared with weak banks. However,one may argue that the former penalise depositors and couldtherefore accuse central banks of failing to deliver growth andfairness.
65 The so-called Pandemic Emergency Purchase Programme(PEPP).
4 TARGET SERVICES AND D-EURO AS A MEANSOF PAYMENT
The present section aims to offer a moredetailed presentation of the design and func-tionalities of a prospective d-euro. Recently,much research activity has been devoted to thedevelopment of CBDC blueprints66 (mainlyAhmat and Bashir 2017; Engert and Fung2017; Bergara and Ponce 2018; Juks 2018;Kumhof and Noone 2018; Panetta 2018;Sveriges Riksbank 2018b; Yao 2018; Armeliuset al. 2018; Norges Bank 2019; Bank of Eng-land 2020; Boar et al. 2020; Jung and Uhlig2020). As a result, many similar theoreticalmodels have been conceived and presented(see also Barrdear and Kumhof 2016; Gouveiaet al. 2017; Engert et al. 2018b; Fernández2018; Mancini-Griffoli et al. 2018; Kahn et al.2019; Bank of England 2020; Bindseil 2020).Such models present several similarities andare often distinguished by subtle differences.The model presented here attributes to theproposed d-euro a number of features that maybe shared by other theoretical efforts. Never-theless, this project is marked by the uniquecharacteristic of trying to implement a CBCDunder the limitations and the design of theEurosystem’s current payment infrastructure.This is the main factor behind some of thespecificities presented here, which will be scru-tinised accordingly and through this prism.
Central banks that operate in the area of pay-ment, clearing and settlement systems have toexamine whether d-euro would improve thesecurity and efficiency of retail payment sys-tems and (in the case of the Eurosystem) TAR-GET services. In fact, central banks, in orderto provide convertibility between differentforms of money, operate RTGS systems (Mar-tin and McAndrews 2008; Kokkola 2010)whereby commercial banks submit paymentorders and settle interbank transactions with-out the use of cash. In addition, such systemsare closely associated with monetary policyconsiderations (TARGET2 settles paymentsrelated to the Eurosystem’s monetary policyoperations and implements monetary policy
decisions) and financial stability (it providessecured liquidity to the financial system). How-ever, when issuing d-euro, central banks haveto apply design features that would allow themto continuously fulfil their mandate of pro-viding safe, stable, efficient, reliable and securepayment systems.67
First and foremost, the d-euro must satisfy thedistinct properties of cash, i.e. to emulate cashin a digital environment. As can be seen in thefollowing table, d-euro is designed as a low-costand efficient means of payment that warrantsmaximum acceptance, accessibility, semi-anonymity, availability, convenience, scalabil-ity, reachability, and stability. In addition, thesupport of monetary policy and the improve-ment of financial stability should be the coreconsiderations of the d-euro design. The twocore principles behind d-euro issuance are: (i) convertibility at par with commercial bankdeposits, which makes it a liability of theEurosystem; and (ii) zero remuneration (likecash), which excludes it from serving as a mon-etary instrument. Moreover, some legal con-siderations related to d-euro’s design should beclarified. Since the Court of Justice of theEuropean Union (CJEU) and the Treaty onthe Functioning of the European Union(TFEU) do not explicitly authorise theEurosystem to issue a CBDC and given that d-euro should have legal tender status (Mancini-Griffoli et al. 2018), a clear authorisation ford-euro issuance does not exist.68 Given that d-euro offers zero remuneration and that a pri-ori its design should improve payment infra-structures’ safety and efficiency, Articles 17 and13.1 of the ESCB/ECB Statute are not rele-vant.69 In addition, it seems that the issuance ofthe d-euro will not affect settlement finalityand insolvency procedures as expressed in the
51Economic BulletinJuly 2020108
66 Some 66 central banks replied to a survey, covering 75% of theworld’s population and 90% of its economic output, and reportedthat more than 80% are working on CBDC (Boar et al. 2020).
67 Under Article 3.1 of the ESCB/ECB Statute.68 Mersch (2020) argued that “without legal tender status, the legal
basis would need to be clarified, as would the relationship betweena CBDC and euro banknotes and coins, along with the process bywhich one could be exchanged for the other”.
69 However, an important legal issue that is outside the scope of thispaper is whether TFEU (for example Article 128) or other legaldocuments need to be amended.
settlement finality directive (SFD) (Article 1 ofSFD). PSD2, although not originally orientedtowards instant payments, assigns central bankswith the tasks of improving the efficiency ofpayments throughout the Monetary Union,ensuring a high level of consumer protection(from fraud), and respecting and protectingfundamental rights, including the ones of pri-vacy and confidentiality (Articles 59, 68, 72-74,94, 97). D-euro compatibility with KYC guide-lines, adherence to AML/CTF requirements,and compliance with the GDPR are legal con-siderations that must also be taken intoaccount, on a preliminary basis, in relation tothe issuance of d-euro (Bank for InternationalSettlements 2015). However, a further in-depthanalysis should establish a robust legal regime.
D-euro is a retail account-based CBDC,whereby central banks verify the validity ofinstant payment transactions by using a cen-tralised ledger structure70. According to thisapproach, households and enterprises do nothave direct access to central bank liabilities butinstead hold accounts with commercial banks.The latter are the administrators of these
51Economic Bulletin
July 2020 109
70 DLT does look very promising. However, at present it is notmature enough to support the issuing of the d-euro. In a previousshort article describing DLT, the author concluded that:“personally, I believe that this new technology could potentiallyprove to be as important as the internet. However, I also believethat we are still at the beginning of the exploration of its promises,while at the same time we need a careful analysis of its possibleconsequences both in terms of monetary policy and financialstability. In fact, the marriage of this new code with the existinginfrastructure of the financial sector has just taken place. And thehoneymoon, exciting and full of inspiration, will be long”(Korfiatis 2018, p. 12).
AcceptanceHigh (majority of retailers still accept cash)
High Low
Access Anyone Anyone Supervised credit institutions
Account provider – Authorised credit institutions Central bank
Anonymity Anonymous Semi-anonymous Not anonymous
Availability 24/7/365 24/7/365 TARGET2 operating hours
ConvenienceMedium (requires both parties to be present)
High High
Convertibility Par Par Par
Cost Medium (withdrawal and storage) Low Low
Finality Final Final and irrevocable Final and irrevocable
Form Physical Digital Electronic
Interest rate No No Yes
Legal tender Yes Yes No
Limits Yes Yes None
Scalability Low and medium value transactions All value transactions Large value transactions
Settlement1 Immediate Instant Real-time
ReachabilityMedium (only proximity payments)
High (pan-European and cross-border)
High (Eurosystem’s supervised creditinstitutions)
Risk Medium (counterfeit, steal) Low None
Cash d-euro Reserves
D-euro as account-based retail central bank digital money
1 See Bech et al. (2017).
accounts and provide any account-related serv-ices. This specific d-euro implementation,whereby commercial banks act as intermedi-aries, requires a public-private partnership ofa manner described later in this section. In fact,the objective of this section is to present themain features of d-euro and its technicaldesign. The table compares d-euro as anaccount-based retail central bank digital cur-rency with the other existing forms of centralbank money.
The features71 of d-euro in a two-dot model:given the proposed attributes shown in thetable above, the designed features of d-euro inthis paper were originally presented by theauthor at the 1st and 2nd 2018 ECBEUROchain Hackathons. Specifically, we pro-pose a two-dot model account-based d-euro.Central banks (ECB and NCBs) are viewed asdot-one entities that settle all d-euro transac-tions by crediting and debiting account bal-ances. End users do not interact directly withcentral banks but indirectly through PaymentService Providers (PSPs). PSPs can be viewedas dot-two entities and mainly function asintermediaries between end users and centralbanks. Commercial banks can act as PSPs andare able to provide their clients with d-euroaccounts. The proposed two-dot structure isexplained in greater detail in the next section.In fact, there are several sets of different pro-posed features that could make CBDC an effi-cient means of payment, a stable unit ofaccount, and a secure store of value. Thatbeing said, it all boils down to a policy decisionwhose adoption depends on the central bank’swish to be involved in the retail payments land-scape and on the implications that this decisionmay have for the safety and the efficiency ofretail payments, monetary policy and financialstability. Given the foregoing argumentation,an ideal account-based d-euro would exhibitthe following features (see also Bank for Inter-national Settlements 2015):
Anonymity: d-euro is semi-anonymous, a prop-erty which renders it the ideal alternative anti-money laundering solution and offers an
advantage in terms of privacy compared withphysical cash72. The attractiveness of this semi-anonymous design rests on the fact thatanonymity may be only lifted due to AML con-cerns and/or suspicion of illegal economicactivity such as money laundering and terror-ist financing. Thus, d-euro satisfies, especiallyfor lower-value payment transactions, the pri-vacy and traceability requirements that helpavert the misuse of the financial system tochannel illicit transactions that damage itsintegrity, stability and reputation.
Availability: d-euro is available to all users, pro-viding a 24/7/365 payment experience andimmediate transfer of funds. Unlike cashwhose availability depends on ATM with-drawals (which are costly and may includewithdrawal fees), visits to a bank branch (whichoperates during banking hours and includestransactional fees) or access to secure storagelocations (which bears security costs), d-eurois universally available, without restrictions,and meets the general trend towards digitali-sation. Thus, it meets users’ need for secureand around-the-clock digital money, irrespec-tive of the underlying payment instrumentused, and can execute both online and offlinepayments.
Convenience: d-euro is convenient, as transac-tions can be initiated via different devices (e.g.mobile phones, wearables and payment cards),channels and technologies (e.g. NFC, QR-code,AI, AR and IoT). In fact, the proposed d-eurohas been designed with convenience in mindand accordingly replicates all the crucial char-acteristics of physical cash (that have made itthe most convenient means of payment so far),while at the same time facilitating the digitaltransformation of the economy (Selgin 1994;Ruth 2018). In addition, with the exception ofweb-based services which are energy-consum-ing and rely on internet access (e.g. smart-
51Economic BulletinJuly 2020110
71 Those characteristics were described and presented at length at the1st and 2nd EUROchain Hackathons (April and November 2018,respectively).
72 The “… CBDC would be acting as an instrument with the liquidityand anonymity of cash, but without the limitations on portabilitythat come with physical cash” (FATF 2020, p. 28).
phones), d-euro can be available to the user viaa payment card,73 thus incorporating the mainfeature of the most popular electronic paymentinstrument, i.e. convenience.
Cost: d-euro is a digital means of paymentalternative to cash and hence its issuance haszero cost on the public. D-euro accounts areprovided, managed and maintained by com-mercial banks, and do not have opening ormaintenance fees. In addition, transferring liq-uidity from a commercial bank deposit accountto a d-euro account is free, while the pricingstructure is the same, i.e. there is free usage ofe-banking or the ATM. Consequently, the pro-posed d-euro is conceived as a cost-efficientmeans of payment that does not suffer fromthe aforementioned cost of cash.
Interest: d-euro offers zero remuneration andaccordingly plays a rather minor role in mon-etary policy implementation and financial sta-bility. Thus, it is attractive to users without vio-lating the parity rule (see further below) andconflicting with anonymity (if interest is sub-ject to taxation). In addition, it discouragescommercial bank account holders from movingliquidity between traditional and d-euroaccounts. Again, with users facing maximumholding limits, it becomes highly unlikely thatthe d-euro would affect the ZLB, since theopportunity for depositors to hold unremu-nerated d-euro as the best costless alternative(avoiding both possible negative bank depositrates and the costs generated to holding cash)is offset by its imposed limit (Dyson and Hodg-son 2016; Yanagawa and Yamaoka 2019).
Issuance: d-euro is issued by central banks inorder to promote convenience, efficiency,transparency, stability and accessibility in retailpayments, and to modernise the Eurosystem’sTARGET services, keeping up with techno-logical developments and consumers’demands. The main advantages of issuing d-euro for central banks, as analysed in the pre-vious sections, are: (i) to regain trust by pro-moting a third form of base money; (ii) torespond to the significant decline in the use of
cash; (iii) to facilitate the digital transforma-tion of the economy by addressing users’ needsand demands for a convenient, secure, global,flexible and instant payment experience; (iv) tocontribute to the strategic autonomy andresilience of European payments, thereby sup-porting the reputation of the Eurosystem andthe international role of the euro; (v) to pro-mote the vision for an innovative retail pay-ment solution providing safety, convenience,cost-efficiency, and pan-European (possiblyglobal) reach; and (vi) to offer an attractivealternative to the proliferation of digital cur-rencies.
Quantity: d-euro holders (households) face anamount limit per account (in the next sectionwe analyse further the amount and the limits).Thus, the maximum amount of d-euro perhousehold is constrained by a limit that can beadjusted in order to safeguard the parity rule.This limit could be subject to change depend-ing on the amount of the banknotes in circu-lation in the euro area (inversely proportional),the financial conditions (in a financial crisiscentral banks could increase the limit) and cer-tain inequalities across EU countries. In thelatter case, wealthier countries may increasethe account’s ceiling as they face a greaterdecline in the use of cash. A household cantransfer liquidity from its traditional to its d-euro account up to the maximum amountdetermined by the central bank. In any case,when there would be a change in the limitamount, central banks have to calculate theexact digital money supply which permits d-euro to function as a stable, reliable, and effi-cient means of payment and to avert a bankrun. Under the model proposed here, enter-prises are allowed to receive d-euro. Such
51Economic Bulletin
July 2020 111
73 The ECB’s European strategy for retail payments includes amarket-led solution that allows users to initiate a payment by usinga card with a common European brand and logo. Recently, theECB has welcomed the decision by 16 European banks to launchthe European Payments Initiative (EPI) as a new Europeanpayment solution. EPI “seeks to replace national schemes for card,online and mobile payments with a unified card and digital walletthat can be used across Europe, thereby doing away with theexisting fragmentation. As it is based on the SCT Inst scheme, itcan immediately capitalise on powerful and sophisticated existinginfrastructures, such as the Eurosystem’s TIPS” (European CentralBank 2020c).
accounts cannot be used for the initiation of d-euro payments. After the receipt of a d-europayment, the funds are irrevocably transferredto the firm’s “traditional” euro bank account.In this way, said digital funds are “subtracted”from circulation and the d-euro money supplyis adjusted accordingly.
Two-dot model: d-euro issuance follows thespecifications of the two-dot model. House-holds and enterprises do not hold d-euroaccounts directly with central banks in order toget access to central banks’ liabilities. At thefirst dot, central banks open Dedicated Digi-tal Accounts (DDAs) in TIPS to TIPS partic-ipants (usually commercial banks already hold-ing DCAs)74. Central banks are responsible forthe opening and management of the accounts,the setting of liquidity limits, the monitoring ofliquidity positions, the performance of localcontingency and the provision of help deskfunctions. At the second dot, commercialbanks do provide end customers with d-euroaccounts. Commercial banks are responsiblefor the opening and management of theaccounts, ensuring adherence to KYC require-ments and providing innovative front-end solu-tions for the initiation of d-euro payments. Infact, we could say that the two-dot model drawsa dividing line between the back-end settle-ment and the front-end customer services, cre-ating an efficient and promising public-privatesector partnership.
Reachability: d-euro transactions are settled inTIPS75 whose pan-European reachability isensured by the Eurosystem.76 TIPS is anextension of TARGET services and enjoys theparticipation of the overwhelming majority ofthe Payment Service Providers (PSPs) in theEEA. Additionally, it is in an optimal positionto facilitate pan-European reachability sinceit had been developed77 to facilitate instantpayments becoming the new normal. There-fore, households and enterprises have the pos-sibility to make and receive domestic andcross-border instant payments in d-euro usingexclusively their central bank d-euro liquiditypool. Users enhance their convenience and
reachability by using IBAN proxies (likemobile phone numbers) based on “trustedhardware”78 and they experience P2P andP2B mobile payment solutions. Moreover,since a multi-currency facility is a core prin-ciple of TIPS,79 every d-euro account holdercould conduct global instant payments via onesingle account in a fast, safe and efficientmanner.
Security: d-euro transactions are settled one byone in central bank money with instant final-ity. Central banks have the statutory task tosecure the smooth operation of payment andsettlement systems. Therefore, settlement incentral bank money eliminates credit risk andensures safety, consistency, efficiency, trans-parency and effectiveness. Since TIPS complieswith all relevant legal and oversight require-ments, it offers a robust and resilient marketinfrastructure that adheres to the Eurosystemoversight requirements and the internationallyapplied CPMI-IOSCO Guidance on cyberresilience. At the front-end, PSPs shall ensurea high level of consumer protection by apply-ing strong customer authentication and com-mon and secure open standards of communi-cation80. Thus, strong cryptography andauthentication would make d-euro the mostsecure digital asset, able to initiate anonymouspayments in central bank money.
Speed: d-euro payments are settled via thestate-of-the-art settlement engine of TIPS that
51Economic BulletinJuly 2020112
74 Strictly legally speaking, TIPS accounts in d-euro are opened inTARGET2 by the competent central bank. In addition, centralbanks, as non-profit organisations, are commercial bank accountprovider agnostic, i.e. every credit institution established in theEuropean Economic Area (EEA) that holds a RTGS DCA incentral bank money may offer d-euro accounts to its customers.
75 The Eurosystem in 2017 offered to settlement banks theconnectivity option of ASI6 RT in order to manage their instantpayment via their respective ACH. However, this solution, whichallows ACHs to offer instant payments in commercial bank moneywith the final settlement in central bank money, requires cashcollateral (which is not included in the calculation of the reserverequirement) in order to remove the counterparty risk associatedwith instant transactions.
76 “TIPS is in an optimal position to facilitate pan-Europeanreachability, but it cannot guarantee it” (European Central Bank2018, p. 3).
77 “The value of TIPS service will be increasing accordingly to thepercentage of participating PSPs” (Korfiatis 2019, p. 37).
78 Trusted hardware encompasses encryption and data authentication.79 Riksbank decided to use TIPS for the Swedish krona (Flodén 2019).80 According to PSD2 and Regulation (EU) 2018/389.
offers final and irrevocable settlement forinstant d-euro payments. A d-euro payment isavailable to the beneficiary in less than ten sec-onds, simulating a cash payment where bothparties are required to be present to settle thetransaction. Although speed in retail paymentshas been recognised as the new normal, recentexperience has showed that both time andeffort must be spent so that end users startadopting instant payments in their daily trans-actions (Hartmann et al. 2019). Payments in d-euro would cost much less than TARGET2customer payments (also suffering from oper-ational time limitations), would not suffer fromfragmentation like the various interbank solu-tions offered by local Automated ClearingHouses (ACH)s81 (usually also operating onlyduring TARGET2 hours) and would thus com-prise one instant, stable, efficient, and user-friendly pan-European payment solution.
The 1:1 parity rule: d-euro supply is perfectlyelastic, as d-euro is issued and exchanged atpar with the other forms of central bank money(Stournaras 2018). Therefore, converting com-mercial bank deposits into d-euro should bethe same as withdrawing cash (in this context,EUR 100 of commercial bank deposits areexchanged for DEUR 100 instead of ban-knotes). Thus, the parity rule ensures that acredit risk-free d-euro could be seen as asecure store of value equal to a commercialbank deposit (since they trade at par withoutentailing any credit risk), as a stable unit ofaccount (since central banks issue the d-euroand commercial banks provide it), and as alow-cost medium of exchange (parity ensuresthat euro and d-euro purchase the same quan-tity of goods and services) (Meaning et al.2018). In other words, d-euro is a stable, secureand easy to use form of money. Additionally,the imposed ceiling on the amount of the d-euro allocated could be seen as adding stabil-ity to the financial market, creating fairness tosociety, and minimising the likelihood of a dig-ital bank run.
Based on these attributes, the next sectionanalyses how the proposed d-euro should be
implemented in order to improve the efficiencyof the system and promote widespread adop-tion and usage. In this context, Figure 3 pres-ents the key benefits of d-euro deriving fromthe aforementioned design features.
Technical design: d-euro issuance follows thestructure of the two-dot model shown in Fig-ure 4. The upper layer of the pyramid is occu-pied by central banks, which are both d-euroissuers and TIPS operators. However, house-holds and enterprises cannot open accountsdirectly with the central bank and consequentlya commercial bank acts as an intermediary toenable their access to central bank money. Inorder to issue d-euro, central banks have tomeet three main objectives: first and foremost,to act as operators, overseers and catalysts ofretail payments systems; second, to set limitson d-euro accounts; and, third, to settle d-eurotransactions in real time via TIPS. While thesethree objectives are complementary, they alsoform three distinct steps of the technicaldesign. Central banks have to convince thepublic to show confidence and trust in d-euroas a settlement asset. Hence, in their capacityas operators, they provide accounts and set-tlement-related services; as overseers, theypromote and ensure the safety and efficiencyof payment systems and instruments; andfinally as catalysts, they prevent market frag-mentation and ensure a level playing field forboth end users and service providers. In short,central banks’ aim will be to create a mean-ingful and secure d-euro ecosystem.
The importance of imposing limits on indi-vidual d-euro holdings and the associated pol-icy implications have already been discussedin this paper. Such individual limits have to bedetermined and adjusted if deemed necessaryby the central banks (Chiu et al. 2019; Bind-seil 2020). These limits are subject to changeaccording to the amount of banknotes in cir-culation, financial market conditions, and cer-
51Economic Bulletin
July 2020 113
81 In Greece, DIAS supports the IRIS Online Payments solution thatenables the immediate execution of interbank fund transfers withinGreece.
51Economic BulletinJuly 2020114
Figure 3 The key benefits of d-eur ro
tnso uymonon ienvennoC
erSecu
Figure 4 The two-dot model
yltantsnistneymaporue-delesndaorue-dseussikanbalrtneC
kanbalrtnecehtn istnuocacADD
era geR
se
smetsysstneymapaneporuEofonotaundaecneliiserehtrt ouppS
ymapalbogl365/7/24rofndamedndaoaislatgiidsdarwotndert’sresuteMe
stneymapilaetrrofhascalcsiyhpuehtnienilcedehttuoabevacorpeB
tneymapofsanemtneicendaelabliestneymaperucesg nomorpy btsurtnaig
payments elaitstrategy for reort the European paymentsppSu
y mo
stnemno
ofes
htiw
ciently nts
ofy
€
sevresermuminimehtn idetalucalctonsiorue-d
snoulsotneyma porue-d
edviorpsPTPndasPSP
stnuocacorue-dg nidloh
snoutsnitiderC
stnuocacorue-d
€€ €
PTR
€
dR
payments elaitstrategy for re
seicnerrucaltigidy tilaerehtotdnopseR
sredlohtnuoccA
stnuocac
tain indices of inequalities across EU coun-tries. The proposed limit has been set up toDEUR 4,000 per active82 person and retiree83.Assuming a euro area population of 342 mil-lion84 people whereby 228 million are activepersons and retirees85, the total amount of ini-tial d-euro supply is estimated at around EUR912 billion. Inactive persons younger than 15years shall not be eligible to hold d-euroaccounts. During episodes of financial crisis,enterprises may be allowed to hold d-euroaccounts (for monetary policy purposes e.g.helicopter money) but in normal circum-stances the default limit is zero since the mainobjective of the d-euro is to maximise theeffectiveness of households’ retail paymenttransactions. The proposed amount ofDEUR 4,000 lies between the value of percapita payments with cards issued by residentPSPs (EUR 5,300) and the value of per capitacash withdrawals at ATMs provided in theeuro area by resident PSPs (EUR 1,110)86.Thus, it is calculated on the basis of the useof cash, which is currently the dominantinstrument in proximity retail payments in theeuro area, and of cards, which form the mostpopular electronic means of retail paymentsat POI in the euro area (46%). In addition,the proposed ceiling of DEUR 4,000 percapita, summing up to EUR 912 billion forthe euro area, seems at present an estimatedamount that covers the basic retail paymentneeds of a household when currency in cir-culation (EUR 1.2 trillion) averages EUR5,300 per active person and retiree, and theaverage monthly net income in the euro areaafter deducting the average monthly savingrate (of 13%)87 amounts to EUR 1,400. How-ever, the above number should be subject toincrease in countries with higher average netincome or if currency in circulation is per-ceived to decline in the euro area (to belowDEUR 912 billion), the reason being thatsome households do not top up their d-euroaccounts or some elderly people use d-euroonly as a store of value.
TIPS provides a harmonised and standard-ised pan-European service that settles
instant payments in central bank money. It isavailable 24/7/365 and hosts euro dedicatedcash accounts (TIPS DCAs) and (in ourmodel) d-euro dedicated digital accounts(TIPS DDAs)88. D-euro would be the secondcurrency settled on the TIPS multi-currencyplatform.89 TIPS DDAs, unlike TIPS DCAs,are not linked to the unique payment accountof TARGET2 (PM account) responsible forthe management of liquidity, and their bal-ance will not count towards the fulfilment ofthe minimum reserve requirements andexcess reserves (Barrdear and Kumhof 2016;Bjerg 2017; Brunnermeier and Niepelt 2019).However, any existing (or future) participantin TARGET2 is seen as an eligible TIPSDDAs holder. Moreover, TIPS allows forthree types of participation:90 (i) participants,i.e. PSPs eligible to open a PM TARGET2account; (ii) reachable parties, i.e. entitiesable to access a participant’s TIPS DCAin order to settle payments in TIPS; and (iii) instructing parties, i.e. entities or thirdparties authorised by participants or reachableparties to send/receive instant payments inTIPS on their behalf. Therefore, it becomesobvious that households can only perform therole of reachable party. Thus, a person whoopens a d-euro account at a commercialbank,91 in order to have access to central bank
51Economic Bulletin
July 2020 115
82 Active population of an economy is the sum of employed andunemployed inhabitants.
83 Population, economic and social conditions data provided byEurostat. Payment data provided by the Statistical Data Warehouse(ECB).
84 https://appsso.eurostat.ec.europa.eu/nui/show.do?dataset=demo_gind&lang=en.
85 The population aged between 15 and 64 is calculated at 160 million,while estimated retirees number 68 million.
86 The distribution of euro banknotes to households is made viaATMs and commercial bank branches.
87 Source: https://ec.europa.eu/eurostat/documents/2995521/10159175/2-14012020-AP-EN.pdf/7d8ef583-6e77-99d5-779e-424de674dd5f.
88 This paper contains an abridged description of how a DDA fits intothe TIPS platform. However, the author, having the experience of beingboth T2S project manager and TIPS project manager for the Bank ofGreece, is more than eager to discuss and/or provide any additionalinformation on the technical aspects of issuing d-euro via TIPS.
89 TIPS could support cross-currency instant payment transactions.However, at the moment, legally speaking, TIPS DDAs, just asTIPS DCAs, fall under the legal and operational perimeter ofTARGET2 as defined in the revised TARGET2 Guideline.
90 In the future, a fourth type of participation may be added, called“Ancillary System”, in order to allow ACHs to interact with TIPS.
91 Participation in TIPS is not mandatory by the Eurosystem.However, participation in TIPS was proposed to be a mandatoryapplication requirement concerning DDA openings by thesignificant entities directly supervised by the ECB.
money, will automatically become a reachableparty to TIPS with a predefined paymentcapacity limit. Local ACHs92 could assume animportant role in achieving settlement ofinstant d-euro payments by acting, based on acontractual agreement, as the instructing partyof a participant. Thus, TIPS seems the opti-mum solution for d-euro settlement, as it pro-vides pan-European reachability,93 settlesinstant payments in central bank money withhigh processing capacity, guarantees securityand around-the-clock availability, and meetsgrowing consumer demand for securedinstant payments with a potential globalacceptance94.
The role of commercial banks in the two-dotmodel is to provide the new ecosystem withthe know-how and best practices that theyhave accumulated from the existing euroecosystem, e.g. the client interface, KYC pro-cedures and front-end innovative paymentsolutions95. On the other hand, householdssubstitute their commercial bank deposits withd-euro ones up to the imposed ceiling. Weassume that, up to a certain estimated limit (inour model the d-euro supply is a policy choiceafter assessing the consequences for monetarypolicy and financial stability), the d-euro willnot significantly affect the size of commercialbanks’ balance sheets but only their compo-sition, as end users substitute euro depositswith d-euro ones. As a result, households’ dis-posable liquidity remains unchanged, whilethey make an important efficiency gain96. Withd-euro replacing only a part of physical cash,we do expect a “stolen benefit”97 effect since,given households’ d-euro limit and incomeconstraint, an increase in d-euro will probablyonly reduce cash withdrawals from ATMs. Cit-izens wishing to open d-euro accounts in orderto initiate instant payments in central bankmoney will have to address to the commercialbanks where they already hold accounts (pay-roll or savings accounts). Commercial banks,according to the KYC guidelines, will opensuch accounts and provide d-euro customersand TPPs with innovative onboarding andfront-end services. Thus, customers will be
able to initiate d-euro payments followingmodern and innovative solutions thanks to theuse of open APIs and user-friendly interfaceseven in a post-PSD2 era. In the default ver-sion, d-euro accounts will be credited or deb-ited during the processing of a retail paymenttransaction or liquidity transfer via TIPS, andP2P payments will be processed anonymously.Anonymity, as well as privacy and data pro-tection are provided by commercial banks. Asalready stated, d-euro is semi-anonymous andtherefore payers do not reveal their identityto the intermediaries. Anonymity is achievedby mapping the account holder’s personalinformation to random unique alphanumericcharacters that are used in the payment mes-sage98. Finally, when a d-euro payment isrequested, commercial banks debit theordering customer’s account, and transfer theinstructed d-euro amount via the central
51Economic BulletinJuly 2020116
92 To date, ACHs cannot open a TIPS account since they have toadhere to the SCT Inst scheme, which is restricted only to PSPs(SCT Inst Rulebook). In addition, it was suggested that ACHscannot open an ASI6 RT technical account for settling d-euro sinced-euro instant payments will strictly be processed within theperimeter of TARGET2.
93 TIPS is an extension of TARGET2 and is based on the SCT Instscheme.
94 TIPS participated in the proof of concept for SWIFT gpi, needingonly 0.06 seconds to process a transaction.
95 For enterprises, commercial banks could support electronicinvoicing presentment and payment (EIPP) solutions both at thedomestic and the European level.
96 Legally, households (acting as reachable parties) own the liquidityof the commercial bank’s DDA (in any other DCA in TARGET2,the owner of the liquidity is the commercial bank). Therefore, thetotal liquidity on the DDA is independent of the main account andcannot be a subject to a change triggered by the TIPS participant(for example to provide a loan).
97 The “stolen benefit” is a term coined by the author in the periodicexhibition of the Bank of Greece’s museum (“e-payments: a roadmap”) to describe the inverse relationship between the use of twosubstitute means of payments. During the Greek crisis,households had to a great extent adapted to conducting electronicpayments. However, this relative increase in card payments (from23,4% in 2015 to 52,6% in 2018 as a percentage of the total numberof payments) was not accompanied by a concurrent decrease in cashwithdrawals from ATMs but by a decline in the use of otherelectronic means of payments (mainly credit transfers).
98 A new global messaging standard, known as “ISO 20022”, althoughit contains the identity of the parties involved in the transaction (thenames of the initiator and of the beneficiary are mandatory messagefields), can facilitate more structured identity fields. Thus, in orderto keep anonymity, the sender’s field is populated with the specialalphanumeric characters provided by the ordering institution. Thisproxy is only known to the latter and the central bank and remainsunknown to the receiver (a cryptographic key is required to openthe message). As a result, only central banks can fully trace thetransfers of funds in order to prevent, investigate and detect moneylaundering and terrorist financing. In specific cases such as taxpayments, it will be possible to use tax identification codes with theinitiator ID instead of the anonymous proxy and hence to removeanonymity.
bank’s platform to the recipient’s d-euroaccount99.
The two-dot model is an efficient approach toconducting d-euro retail payments. It providespan-European reach, and boosts convenience,safety and efficiency by driving central banks,commercial banks, TPPs, and ACHs to offer asuperior payment experience. In this model,the initial use of d-euro as a retail paymentmedium serves as a stepping stone to becomethe new cash of a digitalised ecosystem. Addi-tionally, we have faced a number of challengesconcerning the aforementioned features100. Forexample, due to policy considerations, centralbanks may increase the d-euro limit to avoidcreating a shortage of minimum reserves101.The presented model guarantees a high levelof efficiency and ensures the safety and smoothoperation of payment systems by defining spe-cific roles for all parties involved in the pay-ment chain and the d-euro ecosystem. How-ever, further research is warranted in order tofacilitate its adoption and implementation.
5 CONCLUSION
Τhe widespread use of digital technology hasradically reshaped the way we share andexchange information, as well as how we engagein transactions and perceive payments. Accord-ingly, the ECB has acknowledged that greaterattention should be paid to the front-end sec-tor of European retail payment systems, with aview to protecting its strategic autonomy andresilience and to enhancing the internationalrole of the euro. Nowadays, the ECB has startedweighing the costs and benefits of issuing aCBDC, as well as analysing its feasibility andmerits. In this context, we started our analysisby comparing CBDC with the currently existingforms of money in order to highlight its prop-erty of convertibility, which constitutes the cru-cial building block of our model as it leaves thequantity of central bank money unaffected.
CBDC is the digital cash issued by a centralbank for use in payments and settlement. This
paper presented a retail account-basedCBDC, where the latter is provided indirectlyto the public by commercial banks via com-mercial bank accounts. We argued that thistype of CBDC involves a public-private part-nership whose effectiveness is enhanced byestablishing a separate and distinct role foreach stakeholder.
This paper analysed four main motivations: (i)the issue of trust and the digital transformationof the economy; (ii) the decline in the use ofcash; (iii) financial stability considerations; and(iv) monetary policy considerations. We haveconcluded that central banks have to act proac-tively instead of adopting a defensive stancetowards a downward trend in cash use. D-eurois to play a rather minor role in monetary pol-icy implementation and financial stability con-siderations since it is likely to be irrelevant to
51Economic Bulletin
July 2020 117
99 Assuming that a DEUR 20 payment is initiated. The commercialbank of the payer (Bp) instructs (via TIPS) the central bank to payDEUR 20 (debiting its TIPS DDAp) to the commercial bank of thebeneficiary (Bb) (crediting TIPS DDAb). The central bank debitsthe Bp’s account with DEUR 20 and accordingly the Bp debits thepayer’s account with DEUR 20. Simultaneously, the central bankcredits the Bb’s account with DEUR 20 and accordingly the Bb
credits the beneficiary’s account with DEUR 20. Furthermore, let’sassume that a household triggers a liquidity transfer of EUR 40from its deposit account to its d-euro account. In that case, Bp
orders the central bank to debit its RTGS DCAp (simultaneouslyreducing the available liquidity in its main account and hence itsreserves) with EUR 40 and to credit its TIPS DDAp with DEUR40. Therefore, while the household’s liquidity remains the same,there are only a change in the composition of central bank moneyand a decrease in the commercial bank’s reserves.
100 Paraphrasing Lancastrian demand theory, the two-dot model hasbeen carefully designed by predicting demand on the basis ofcertain d-euro features (attributes) in order to make it households’preferable means of payment in a retail payment. Deciding on thedesirability of individual d-euro features, our motivation has beento achieve widespread adoption and hence to create strongnetwork effects and boost the effectiveness of the system. One mayargue that proposing a low limit is not the most desirable feature,compared with a private issuer (with no imposed limits) or SEPAInst (with a limit of up to EUR 100,000), since some users mayfind d-euro restrictive in their desired purchases. However, in ourmodel, limits are subject to change according to the foregoingargumentation. In addition, there is a possibility for a certain limitto the provision of full anonymity (the proposed amount was EUR500 following the Greek ceiling on cash payments), as presentedby the author at the 2nd EUROchain Hackathon.
101 The ECB requires commercial banks established in the euro areato hold minimum reserves. However, when central banks increasethe d-euro supply (and assuming that households top up their d-euro account up to the new ceiling), a shortage of liquidity(reserves) is created. This happens because households transferliquidity from deposits to d-euro accounts (both held with the samecommercial bank). This conversion of commercial bank money intocentral bank money is mapped by a transfer from the TARGETPM account (Main Cash Account – MCA) after the T2-T2Sconsolidation) to the TIPS DDA. In normal financial conditions,central banks use open market operations (OMO) to supplycommercial banks with the necessary liquidity (increasing reserves).
inflation targeting. However, it may be used asa direct channel through which central bankscan transmit monetary policy decisions tohouseholds and enterprises.
An account-based d-euro has to follow six prin-ciples: (i) semi-anonymity (d-euro should besemi-anonymous due to AML concerns); (ii)trust (digital trust is issued directly by centralbanks); (iii) scalability (it should form the cor-nerstone of the efforts to respond to moderntechnological challenges, as well as to enhancethe position of the euro as an international cur-rency); (iv) interest rate neutrality (it should beinterest rate-free and hence unfitting as a mon-etary policy tool); (v) acceptance (it should beused by payment solutions that offer pan-Euro-pean reachability and possible global accept-ance); and (vi) convenience (it should be tailor-made according to customers’ demands for con-venience, speed, and efficiency). We arguedthat the aforementioned principles also con-stitute the key benefits of the d-euro.
When issuing d-euro, central banks have toconsider how to fulfil their mandate of pro-viding safe, stable, efficient, reliable and securepayment systems. We presented a framework tostudy the specific attributes of an ideal account-based d-euro and we proposed the technicaldesign of a two-dot model. Additionally, weshowed how d-euro should be implemented inorder to improve the efficiency of the paymentinfrastructure and to promote its widespreadadoption and usage. That being said, given theintroduced technical design, we discussed theimportance of imposing limits on individual d-euro holdings, as well as the associated policyimplications. Specifically, we proposed a con-crete non-binding limit that is subject to changedepending on the amount of banknotes in cir-culation, financial market conditions, and cer-tain regional inequality indices.
We proposed TIPS as the harmonised andstandardised pan-European service that settlesinstant payments in d-euro. TIPS is available
24/7/365 and d-euro would be the second cur-rency settled on its multi-currency platform,where commercial banks would hold d-eurodedicated digital accounts (TIPS DDAs), andhouseholds may connect too as reachable par-ties. We have concluded that, as the presentDLT technology is not mature enough, TIPSseems the optimum solution for settlement ind-euro, as it provides pan-European reacha-bility, settlement in central bank money, highprocessing capacity, security and around-the-clock availability, and meets growing consumerdemand for secured instant payments andglobal acceptance.
Central banks offer the back-end solutions andare the guarantors of anonymity. On the otherhand, the role of commercial banks is to pro-vide the new ecosystem with their know-howand best practices. Our model clarifies the pro-vision of anonymity, as well as of privacy anddata protection. We have concluded that thetwo-dot model is an efficient approach to con-ducting d-euro retail payments. It providespan-European reach, and boosts convenienceand safety by driving all participating stake-holders (central and commercial banks,TPPs, and ACHs) to offer a superior paymentexperience. According to our understanding, d-euro could be used initially as a retail paymentmedium and serve as a stepping stone towardsbecoming the new digital cash.
Central banks have been responding to thefact that digital currencies are part of theglobal monetary system. Thus it is in theirinterest to ensure that they are not left behindin the technology race. Whether or not d-eurowill be the next step in the evolution of moneyis subject to further analysis. Issuance of sucha currency should contribute towards theautonomy and the resilience of the Europeanpayment ecosystem as well as a climate-neu-tral and digital Europe. This paper argues thatd-euro may be the answer to central banks’concerns about the issuance of a digitalmedium of payment.
51Economic BulletinJuly 2020118
Adrian, T. and T. Mancini-Griffoli (2019), “The Rise of Digital Money”, IMF FinTech Notes,Note 19/01.
Agarwal, R. and M. Kimball (2015), “Breaking through the Zero Lower Bound”, IMF WorkingPaper No. 224.
Agarwal, R. and M. Kimball (2019), “Enabling Deep Negative Rates to Fight Recessions: AGuide”, IMF Working Paper No. 84.
Agur, I. (2018), “Central Bank Digital Currencies: An Overview of Pros and Cons”, in: E. Gnanand D. Masciandaro (eds.), Do We Need Central Bank Digital Currency?, SUERF ConferenceProceedings, 113-122.
Agur, I., A. Anil and G. Dell’Arricia (2019), “Designing Central Bank Digital Currencies”, IMFWorking Paper No. 252.
Ahmat, N. and S. Bashir (2017), “Central Bank Digital Currency: A Monetary Perspective”, StaffInsights, Bank Negara Malaysia, No. 11.
Akerlof, G. (1970), “The Market for ‘Lemons’: Quality Uncertainty and the Market Mechanism”,The Quarterly Journal of Economics, Vol. 84(3), 488-500.
Albertazzi, U., F. Barbiero, D. Marques-Ibanez, A. Popov, C.R. D’Acri and T. Vlassopoulos(2020), “Monetary Policy and Bank Stability: The Analytical Toolbox Reviewed”, EuropeanCentral Bank, Working Paper Series, No. 2377.
Ali, R., J. Barrdear, R. Clews and J. Southgate (2014a), “Innovations in Payment Technologiesand the Emergence of Digital Currencies”, Bank of England, Bank of England Quarterly Bul-letin, Q3, 262-75.
Ali, R., J. Barrdear, R. Clews and J. Southgate (2014b), “The Economics of Digital Currencies”,Bank of England, Bank of England Quarterly Bulletin, 54(3), 276-86.
Allen, H. and A. Dent (2010), “Managing the Circulation of Banknotes”, Bank of England, Bankof England Quarterly Bulletin, Q4.
Altavilla,C., L. Burlon, M. Giannetti and S. Holton (2019), “Is there a zero lower bound? Theeffects of negative policy rates on banks and firms”, European Central Bank, Working PaperSeries, No. 2289, June.
Andolfatto, D. (2015), “Fedcoin: on the Desirability of a Government Cryptocurrency”, Macro-Mania, Blogpost, February.
Andolfatto, D. (2018), “Assessing the Impact of Central Bank Digital Currency on Private Banks”,Federal Reserve Bank of St. Louis Working Paper No. B26.
Arango, C., D. Hogg and A. Lee (2015) “Why Is Cash (Still) so Entrenched? Insights from Cana-dian Shopping Diaries”, Contemporary Economic Policy, 33, 141-58.
Armelius, H., P. Boel, C.A. Claussen and M. Nessen (2018), “The E-krona and the Macro-economy,” Sveriges Riksbank, Sveriges Riksbank Economic Review, No. 3, 43-65.
Athey, S., C. Catalini and C. Tucker (2017), “The Digital Privacy Paradox: Small Money, SmallCosts, Small Talk”, NBER Working Papers, No. 23488.
Auer, R. (2019), “Beyond the Doomsday Economics of ‘Proof-of-Work’ in Cryptocurrencies”,BIS Working Papers, No. 765.
Bank of England (2020), “Central Bank Digital Currency: Opportunities, Challenges and Design”,Discussion Paper, March.
Bank for International Settlements (2003), “The Role of Central Bank Money in Payment Sys-tems”, August.
Bank for International Settlements (2015), “Digital Currencies”, November.Bank for International Settlements (2018), “Central Bank Digital Currencies”, March.Barontini, C. and H. Holden (2019), “Proceeding with Caution – A Survey on Central Bank Dig-
ital Currency”, BIS Papers, No. 101.
51Economic Bulletin
July 2020 119
R E F E R ENC E S
Barrdear, J. and M. Kumhof (2016), “The Macroeconomics of Central Bank Issued Digital Cur-rencies”, Bank of England, Bank of England Working Papers, No. 605.
Bech, M.L. and R. Garratt (2017), “Central Bank Cryptocurrencies”, BIS Quarterly Review, Sep-tember.
Bech, M., Y. Shimizu and P. Wong (2017), “The Quest for Speed in Payments”, BIS QuarterlyReview, 57-68, March.
Bech, M.L., U. Faruqui, F. Ougaard and C. Picillo (2018), “Payments are A-Changin’ but CashStill Rules”, BIS Quarterly Review, 67-80, March.
Bergara, M. and J. Ponce (2018), “Central Bank Digital Currency: The Uruguayan E-Peso Case”,in: E. Gnan and D. Masciandaro (eds.), Do We Need Central Bank Digital Currency?, SUERFConference Proceedings, 82-90.
Biais, B., C. Bisière, M. Bouvard and C. Casamatta (2019),“The Blockchain Folk Theorem”,Review of Financial Studies, 32(5), 1662-1715.
Bindseil, U. (2020), “Tiered CBDC and the Financial System”, European Central Bank, Work-ing Paper Series, No. 2351.
Bjerg, O. (2017), “Designing New Money - The Policy Trilemma of Central Bank Digital Cur-rency”, Copenhagen Business School, CBS Working Papers, June.
Bjerg, O. (2018), “Breaking the Gilt Standard: The Problem of Parity in Kumhof and Noone’sDesign Principles for Central Bank Digital Currencies”, Copenhagen Business School, CBSWorking Papers, August.
Boar, C., H. Holden and A. Wadsworth (2020), “Impending Arrival – A Sequel to the Surveyon Central Bank Digital Currency”, BIS Paper No. 107.
Böhme, R., N. Christin, B. Edelman and T. Moore (2015), “Bitcoin: Economics, Technology,and Governance”, Journal of Economic Perspectives, 29(2), 213-238.
Bordo, M.D. and A.T. Levin (2017), “Central Bank Digital Currency and the Future of Mone-tary Policy”, NBER Working Paper Series, No. 23711.
Bordo, M.D. and A.T. Levin (2018), “Central Bank Digital Cash: Principles and Practical Steps”,in: E. Gnan and D. Masciandaro (eds.), Do We Need Central Bank Digital Currency?, SUERFConference Proceedings, 125-142.
Borgonovo, E., S. Caselli, A. Cillo and D. Masciandaro (2018), “Between Cash, Deposit And Bit-coin: Would We Like A Central Bank Digital Currency? Money Demand And ExperimentalEconomics”, BAFFI CAREFIN Working Papers, No. 1875.
Bossone, B. (2001), “Should Banks be Narrowed?”, IMF Working Paper No. 159.Broadbent, B. (2016), “Central Banks and Digital Currencies”, speech at the London School of
Economics, March.Brown, M., I. Evangelou and H. Stix (2017), “Banking Crisis-Bail-Ins and Money Holdings”, Cen-
tral Bank of Cyprus Working Paper No. 2.Brunnermeier, M.K. and H. James (2019), “The Digitalization of Money”, Princeton University,Working Papers, August.
Brunnermeier, M.K. and D. Niepelt (2019), “On the Equivalence of Private and Public Money”,NBER Working Paper Series, No. 25877.
Buiter, W. H. (2016), “The Simple Analytics of Helicopter Money: Why it Works - Always”, Eco-nomics, Vol. 8.
Bullmann, D., A. Klemm and A. Pinna (2019), “In Search for Stability in Crypto-Assets: Are Sta-blecoins the Solution?”, ECB Occasional Paper No. 230.
Carstens, A. (2019), “The Future of Money and Payments”, speech held in Dublin, March.Chaum, D. (1983), “Blind Signatures for Untraceable Payments”, Advances in Cryptology,
Springer, 199-203.
51Economic BulletinJuly 2020120
Chimienti, M.T., U. Kochanska and A. Pinna (2019), “Understanding the Crypto-asset Phe-nomenon, its Risks and Measurement issues”, European Central Bank, ECB Economic Bul-letin, Issue 5/2019.
Chiu, J. and T.-N. Wong (2014) “E-money: Efficiency, Stability and Optimal Policy”, Bank ofCanada Staff Working Paper No. 16, April.
Chiu, J. and T.-N. Wong (2015), “On the Essentiality of E-Money”, Bank of Canada Staff Work-ing Paper No. 4.
Chiu, J., M. Davoodalhosseini, J.H. Jiang and Y. Zhu (2019), “Central Bank Digital Currencyand Banking”, Bank of Canada Staff Working Paper No. 20.
Cœuré, B. (2018), “The Future of Central Bank Money”, Speech held in Geneva, May.Cœuré, B. (2019a), “Digital Challenges to the International Monetary and Financial System”,
Speech held in Luxembourg, September.Cœuré, B. (2019b), “Monetary Policy: Lifting the Veil of Effectiveness”, Speech held in Frank-
furt, December.Cœuré, B. (2019c), “Towards the Retail Payment of Tomorrow: A European Strategy”, Speech
held in Frankfurt, November.Cruijsen van der, C., L. Hernández and N. Jonker (2015), “In Love with the Debit Card but Still
Married to Cash”, DNB Working Paper No. 461.Davoodalhosseini, M. and R. Mohammad (2018), “Central Bank Digital Currency and Mone-
tary Policy”, Bank of Canada Staff Working Paper No. 36.Dell’Ariccia, G., V. Haksan, T. Mancini-Griffoli, K. Eckhold, S. Gray, F. Han, G.H. Hong, E.
Lundback, H. Oura, H. Poirson, P. Rabanal and D. Sandri (2017), “Negative Interest Rate Poli-cies – Initial Experiences and Assessments”, IMF Policy Papers, August.
Dyson, B. and G. Hodgson (2016), “Digital Cash: Why Central Banks Should Start Issuing Elec-tronic Money,” Positive Money.
Edgar, F. (2012), “The Myth of the “Cashless Society”: How much of America’s currency is over-seas?” Munich Personal RePEc Archive No. 42169.
Eggertsson, G. and M. Woodford (2003), “The Zero Bound on Interest Rates and Optimal Mon-etary Policy”, Brookings Papers on Economic Activity, Vol. 34, 139-233.
Engert, W. and B. Fung (2017), “Central Bank Digital Currency: Motivations and Implications”,Bank of Canada Staff Discussion Paper No. 16.
Engert, W. and B. Fung (2020), “A Uniform Currency in a Cashless Economy”, Bank of Canada,Staff Analytical Notes, March.
Engert, W., B. Fung and S. Hendry (2018a), “Is a Cashless Society Problematic?”, Bank of CanadaStaff Discussion Paper No. 12.
Engert, W., B. Fung and S. Hendry (2018b), “Motivations and Implications of a Central BankDigital Currency”, in: E. Gnan and D. Masciandaro (eds.), Do We Need Central Bank Digi-tal Currency?, SUERF Conference Proceedings, 56-70.
Esselink, H. and L. Hernández (2017), “The Use of Cash by Households in the Euro Area”, Euro-pean Central Bank, Occasional Paper Series, No. 201.
European Central Bank (2015), “Virtual Currency Schemes – A Further Analysis”, February.European Central Bank (2018), “Co-existence of TIPS with other Instant Payment Services”,https://www.ecb.europa.eu/paym/pdf/consultations/tips-coexistence_of_tips_with_other_instant_payment_services.pdf?eaefb18ef7ff74a6c67b357d95a34aac.
European Central Bank (2019a), “Crypto-Assets: Implications for Financial Stability, MonetaryPolicy, and Payments and Market Infrastructures”, ECB Occasional Paper Series, No. 223, May.
European Central Bank (2019b), “Exploring Anonymity in Central Bank Digital Currencies”,In Focus, No. 4, December.
51Economic Bulletin
July 2020 121
European Central Bank (2019c), “Stablecoins – No Coins, but are They Stable?”, In Focus, No.3, November.
European Central Bank (2020a), “Electronic Money”, accessed March 2020, https://www.ecb.europa.eu/stats/money_credit_banking/electronic_money/html/index.en.html.
European Central Bank (2020b), “ECB announces €750 billion Pandemic Emergency PurchaseProgramme (PEPP)”, https://www.ecb.europa.eu/press/pr/date/2020/html/ ecb.pr200318_1~3949d6f266.en.html accessed March 2020, https://www.ecb.europa.eu/press/pr/date/2020/html/ecb.pr200318_1~3949d6f266.en.html.
European Central Bank (2020c), “ECB Welcomes Initiative to Launch New European PaymentSolution”, https://www.ecb.europa.eu/press/pr/date/2020/html/ecb.pr200702~214c52c76b.en.html, June.
Fatás, A. (ed.) (2019), The Economics of Fintech and Digital Currencies, VoxEU.org, eBook.Fernández de Lis, S. (2018), “Central Bank Digital Currencies: Features, Options, Pros and Cons”,
in: E. Gnan and D. Masciandaro (eds.), Do We Need Central Bank Digital Currency?, SUERFConference Proceedings, 46-55.
Financial Action Task Force (FATF) (2020), “FATF Report to the G20 Finance Ministers andCentral Bank Governors on So-called Stablecoins”, June.
Fleuret, F. and T. Lyons (2020), “Blockchain and the Future of Digital Assets”, European UnionBlockchain Observatory and Forum, February.
Flodén, M. (2019), “Should the Riksbank’s Payment System be Open 24/365?”, speech held inStockholm, June.
Friedman, C. (2000), “Monetary Policy Implementation: Past, Present and Future - Will Elec-tronic Money Lead to the Eventual Demise of Central Banking?”, International Finance, Vol.3, 221-27.
Fung, B. and H. Halaburda (2014), “Understanding Platform-Based Digital Currencies”, Bankof Canada, Bank of Canada Review, 12-20.
Fung, B. and H. Halaburda (2016), “Central Bank Digital Currencies: A Framework for Assess-ing Why and How”, Bank of Canada, Discussion Papers, No. 22.
G7 Working Group on Stablecoins (2019), “Investigating the Impact of Global Stablecoins”, Octo-ber.
Gandal, N. and H. Halaburda (2014), “Competition in the Cryptocurrency Market”, Bank ofCanada Working Paper No. 33, August.
Gans, J. and H. Halaburda (2013), “Some Economics of Private Digital Currency”, Bank ofCanada Working Paper No. 38, November.
Garratt, R. and M. van Oordt (2019), “Privacy as a Public Good: A Case for Electronic Cash”,Bank of Canada, Staff Working Papers, No. 24.
Giffen, R. (1891), “The Gresham Law”, The Economic Journal, Vol. 1, 304-306.Gnan, E. and D. Masciandaro (2018), “Do We Need Central Bank Digital Currency? Econom-
ics, Technology and Institutions”, in: E. Gnan and D. Masciandaro (eds.), Do We Need Cen-tral Bank Digital Currency?, SUERF Conference Proceedings, 8-21.
Goodfriend, M. (2016), “The Case for Unencumbering Interest Rate Policy at the Zero LowerBound”, Jackson Hole Conference, August.
Gouveia, O., E. Dos Santos, S. Fernández de Lis, A. Neut and J. Sebastián (2017), “Central BankDigital Currencies, Assessing Implementation Possibilities and Impacts”, BBVA Working PaperNo. 04.
Grym, A., P. Heikkinen, K. Kauko and K. Takaka (2017), “Central Bank Digital Currency”, Bankof Finland, Bank of Finland Economics Review, No. 5.
Habermeier, K., L. Jacome, T. Mancini-Griffoli, C. Baba, T. Bayoumi, J. Chen, G. Dell’Aric-cia, S. Gray, M. Ismael, T. Mondino, A. Pescatori, T.S. Sedik, S. Sgherri, H. Tanimoto, K. Ueda,
51Economic BulletinJuly 2020122
N. Valckx and F. Valencia (2018), “Unconventional Monetary Policies - Recent Experiencesand Prospects”, IMF Policy Papers, April.
Hall, R. (2013), “The routes into and out of the Zero Lower Bound”, paper presented at theKansas City Federal Reserve Economic Policy Symposium, August.
Hamilton, J.D. and J.C. Wu (2011), “The Effectiveness of Alternative Monetary Policy in a ZeroLower Bound Environment”, National Bureau of Economic Research, Report No. 16956.
Hartmann, M., L. Hernandez-van Gijsel, M. Plooij and Q. Vandeweyer (2019), “Are instant pay-ments becoming the new normal? A comparative study”, European Central Bank, OccasionalPaper Series, No. 229.
Hasan, I., T. de Renzis and H. Schmiedel (2013), “Retail Payments and the Real Economy”, Euro-pean Central Bank Working Paper No. 1572.
He, D., R.B. Leckow, V. Haksar, T. Mancini-Griffoli, N. Jenkinson, M. Kashima, T. Khiaonarong,C. Rochon and H. Tourpe (2017), “Fintech and Financial Services; Initial Considerations”, IMFStaff Discussion Note No. 05.
Jobst, C. and H. Stix (2017), “Doomed to Disappear: The Surprising Return of Cash across Timeand across Countries”, CEPR Discussion Paper No. 12327.
Judson, R. (2018), “Big Note, Small Note: Central Bank Digital Currency and Cash”, in: E. Gnanand D. Masciandaro (eds.), Do We Need Central Bank Digital Currency?, SUERF ConferenceProceedings, 23-32.
Juks, R. (2018), “When a Central Bank Digital Currency Meets Private Money: Effects of an E-krona on Banks”, Sveriges Riksbank, Sveriges Riksbank Economic Review, No. 3, 79-98.
Jung, A. and H. Uhlig (2019), “Monetary Policy Shocks and the Health of Banks”, European Cen-tral Bank, Working Paper Series, No. 2303.
Kahn, C. and W. Roberds (2009), “Why pay? An Introduction to Payments Economics”, Jour-nal of Financial Intermediation, No. 3, 1-23.
Kahn, C.M., F. Rivadeneyra and T.N. Wong (2019), “Should the Central Bank Issue E-Money?”Federal Reserve Bank of St. Louis, Working Papers, No. 3.
Kashyap, A., R. Rajan and J. Stein (2002), “Banks as Liquidity Providers: An Explanation forthe Coexistence of Lending and Deposit-Taking”, The Journal of Finance, No. 57, 33-73.
Keister, T. and D.R. Sanches (2019), “Should Central Banks Issue Digital Currency?” FederalReserve Bank of Philadelphia, Working Papers, No. 26.
Kim, Y.S. and O. Kwon (2019), “Central Bank Digital Currency and Financial Stability”, Bankof Korea, Economic Research Institute, Working Papers, No. 6.
Kiyotaki, N. and R. Wright (1989), “On Money as a Medium of Exchange”, The University ofChicago Press, Journal of Political Economy, No. 97, 927-954.
Klee, E. (2008), “How People Pay: Evidence from Grocery Store Data”, Journal of MonetaryEconomics, 55, 526-41.
Kokkola, T. (2010), Payments, Securities and Derivatives, and the Role of the Eurosystem, Euro-pean Central Bank.
Korfiatis, Y. (2018), “DLT and CBDC: A Long Honeymoon”, Kyklos, No. 26, Spring, 30-31 (in Greek).Korfiatis, Y. (2019), “Instant Payments bring the Revolution to Greece and Europe”, Econo-mia, No. 982, 36-37, (in Greek).
Kumhof, M. and C. Noone (2018), “Central Bank Digital Currencies – Design Principles and Bal-ance Sheet Implications”, Bank of England Staff Working Paper No. 725.
Lagarde, C. (2018), “Winds of Change: The Case for a New Digital Currency”, speech held inSingapore Fintech Festival, Singapore.
Lenza, M., H. Pill and L. Reichlin (2010), “Monetary Policy in Exceptional Times”, EuropeanCentral Bank, Working Paper Series, No. 1253.
Libra (2019), “White Paper”, accessed on 11 March 2020, <https://libra.org/en-US/white-paper/>.
51Economic Bulletin
July 2020 123
Mancini-Griffoli, T. and A. Ranaldo (2011), “Limits to Arbitrage During the Crisis: Funding Liq-uidity Constraints and Covered Interest Parity”, SSRN.
Mancini-Griffoli, T., M. Martinez Peria, I. Agur, A. Ari, J. Kiff, A. Popescu and C. Rochon (2018),“Casting Light on Central Bank Digital Currencies”, IMF Staff Discussion Note No. 08.
Martin, A. and J. McAndrews (2008), “An Economic Analysis of Liquidity-Saving Mechanisms”,Federal Reserve Bank of New York, Economic Policy Review, Vol. 14, 25-39.
Masciandaro, D. (2018), “The Demand for a Central Bank Digital Currency: A Political Econ-omy Approach”, in: E. Gnan and D. Masciandaro (eds.), Do We Need Central Bank DigitalCurrency?, SUERF Conference Proceedings, 72-80.
McAndrews, J. (2015), “Negative Nominal Central Bank Policy Rates – Where Is the LowerBound?”, Remarks at the University of Wisconsin, May.
McLeay, M., A. Radia and R. Thomas (2014), “Money in the Modern Economy: An Introduc-tion”, Bank of England, Bank of England Quarterly Bulletin, Q1.
Meaning, J., B. Dyson, J. Barker and E. Clayton (2018), “Broadening Narrow Money: MonetaryPolicy with a Central Bank Digital Currency”, Bank of England Working Paper No. 724.
Mersch, Y. (2020), “An ECB digital currency – a flight of fancy?”, speech held in Frankfurt atthe Consensus 2020 Virtual Conference, May.
Middleton, P. and J. Lund (2018), “Central Bank Digital Currencies”, OMFIF – IBM.Middleton, P. and S. Sinha (2019), “Retail CBDCs: The Next Payment Frontier”, OMFIF – IBM. Milne, A.K.L. (2018), “Argument by False Analogy: The Mistaken Classification of Bitcoin as
Token Money”, Loughborough University.Nagel, E. (1963), “Assumptions in Economic Theory”, The American Economic Review, 53(2),
211-219.Nakamoto, S. (2008), “Bitcoin: A Peer-to-Peer Electronic Cash System”, Bitcoin Org.Narayanan, A., B. Bonneau, E. Felten, A. Miller, and S. Goldfeder (2016), Bitcoin and cryp-tocurrency technologies: A comprehensive introduction, Princeton: Princeton University Press.
Niepelt, D. (2018), “Central Bank Digital Currency: What Difference does it Make?”, in: E. Gnanand D. Masciandaro (eds.), Do We Need Central Bank Digital Currency?, SUERF ConferenceProceedings, 99-112.
Niepelt, D. (2019), “Reserves for All? Central Bank Digital Currency, Deposits, and their (Non)-Equivalence”, CESifo Working Paper Series, No. 7176.
Norges Bank (2019), “Central Bank Digital Currencies”, Norges Bank Papers, No. 1.Panetta, F. (2018), “21st Century Cash: Central Banking, Technological Innovation and Digi-
tal Currency”, in: E. Gnan and D. Masciandaro (eds.), Do We Need Central Bank Digital Cur-rency?, SUERF Conference Proceedings, 23-32.
Pennacchi, G. (2012), “Narrow Banking”, Annual Review of Financial Economics, Vol. 4, 141-159.
Pichler, P. and M. Summer (2018), “Digital Money, Cryptocurrency and Central Banks”, in: E.Gnan and D. Masciandaro (eds.), Do We Need Central Bank Digital Currency?, SUERF Con-ference Proceedings, 91-98.
Quinn, S. and W. Roberds (2014), “How Amsterdam got Fiat Money”, Journal of Monetary Eco-nomics, Vol. 66, 1-12.
Raskin, M. and D. Yermack (2016), “Digital Currencies, Decentralised Ledgers, and the Futureof Central Banking”, NBER Working Paper No. 22238.
Reichlin, L., A. Turner and M. Woodford (2013), “Helicopter Money as a Policy Option”, VOXCERP Policy Portal, May.
Robleh, A., J. Barrdear, R. Clews and J. Southgate (2014), “Innovations in Payment Technologiesand the Emergence of Digital Currencies”, Bank of England, Bank of England Quarterly Bul-letin, Vol. 54, 262-75, September.
51Economic BulletinJuly 2020124
Rogoff, K.S. (2014), “Costs and Benefits to Phasing out Paper Currency”, NBER Working PaperNo. 20126.
Rogoff, K.S. (2016), The Curse of Cash, Princeton University Press.Ruth, J. (2018), “Big Note, Small Note: Central Bank Digital Currency and Cash”, in: E. Gnan
and D. Masciandaro (eds.), Do We Need Central Bank Digital Currency?, SUERF ConferenceProceedings, 34-44.
Selgin, G. (1994), “On Ensuring the Acceptability of a New Fiat Money”, Journal of Money,Credit and Banking, No. 26, 808-26.
Stein, J.C. (2012), “Monetary Policy as Financial Stability Regulation”, The Quarterly Journalof Economics, 127(1), 57-91.
Stournaras, Y. (2018), “The future of money”, Speech held in Athens, May.Sveriges Riksbank (2017), “Central Bank Digital Currencies”, Technical Report on the E-kronaProject.
Sveriges Riksbank (2018a), “Special Issue on the E-krona”, Sveriges Riksbank Economic Review.Sveriges Riksbank (2018b), “The Riksbank’s E-krona Project”, Technical Report, No. 2.Sveriges Riksbank (2019), “The Riksbank Proposes a Review of the Concept of Legal Tender”,
Press Release, April.Tobias, A. (2019), “Stablecoins, Central Bank Digital Currencies, and Cross-Border Payments:
A New Look at the International Monetary System”, Remarks held in Zurich, May.Tobias, A. and T. Mancini-Griffoli (2019), “The Rise of Digital Money”, International Mone-
tary Fund, IMF Note/19/01.Wakamori, N. and A. Welte (2017), “Why Do Shoppers Use Cash? Evidence from Shopping Diary
Data”, Journal of Money, Credit and Banking, 49(1), 115-169.Ward, O. and S. Rochemont (2019), “Understanding Central Bank Digital Currencies (CBDC)”,
Institute of Faculty and Affairs, March.Weber, W. (2015), “Government and Private e-Money-Like Systems: Federal Reserve Notes and
National Bank Notes”, Bank of Canada Working Paper No. 18, June.World Economic Forum (2020), “Central Bank Digital Currency Policy – Maker Toolkit”, Jan-
uary.Wright, R., E. Tekin, V. Topalli, C. McClellan, T. Dickinson and R. Rosenfeld (2017), “Less Cash,
Less Crime: Evidence from the Electronic Benefit Transfer Program”, Journal of Law and Eco-nomics, 60(2), 361-383.
Yanagawa, N. and H. Yamaoka (2019), “Digital Innovation, Data Revolution and Central BankDigital Currency”, Bank of Japan Working Paper Series, No. 19-E-2.
Yang, B.Z. (2007), “What is (Not) Money? Medium of Exchange ≠ Means of Payment”, TheAmerican Economist, No. 51, 101-104.
Yao, Q. (2018), “A Systematic Framework to Understand Central Bank Digital Currency”, Sci-ence China Information Sciences, No. 033101.
Yermack, D. (2013), “Is Bitcoin a Real Currency? An Economic Appraisal”, NBER WorkingPaper No. 19747.
51Economic Bulletin
July 2020 125
51Economic BulletinJuly 2020126
51Economic Bulletin
July 2020
277. Perceived vs actual financial crisis andbank credit standards: is there any indi-cation of self-fulfilling prophecy?Dimitrios Anastasiou, ZachariasBragoudakis and Stelios Giannoulakis
278. The effect of Emergency Liquidity Assis-tance (ELA) on bank lending during theeuro area crisisHeather D. Gibson, Stephen G. Hall,Pavlos Petroulas, Vassilis Spiliotopoulosand George S. Tavlas
279. On the controversy over the origins ofthe Chicago Plan for 100 percentreserves George S. Tavlas
280. A study of the effect of data transfor-mation and “linearization” on time seriesforecasts. A practical approachAlexandros E. Milionis and Nikolaos G.Galanopoulos
281. Did the absence of a central bank back-stop in the sovereign bond markets exac-erbate spillovers during the euro areacrisis? Heather D. Gibson, Stephen G. Hall,Deborah GeFang, Pavlos Petroulas andGeorge S. Tavlas
282. A suggestion for a Dynamic Multi Fac-tor Model (DMFM)Heather D. Gibson, Stephen G. Hall andGeorge S. Tavlas
This section contains the abstracts of Working Papers authored by Bank of Greece staff and/orexternal authors and published by the Bank of Greece. The unabridged version of these texts isavailable on the Bank of Greece’s website (www.bankofgreece.gr).
CONTENTS
WORK I NG P A P ER S(JANUARY – JULY 2020)
127
The authors link senior bank loan officers’responses regarding their decisions for bankcredit standards, from successive rounds ofthe European Bank Lending Survey, to inves-tigate two important issues. First, they exam-ine the relationship between bank credit stan-dards (CS) and perceived and actual financialcrisis. Second, they investigate whether thenotion of the self-fulfilling prophecy is appli-cable in the case of the 2008 global financialcrisis. In particular, the second main researchquestion that they try to answer is whether the
perceived crisis (as implied by the Googlesearch query “financial crisis”) contributed tothe acceleration of the outburst of the actualcrisis. The authors find that both perceivedand actual financial crisis affect senior bankloan officers’ credit standards, with the actualcrisis having the greatest impact. These resultsare consistent both in the short and in the longrun. Finally, by putting forward a binarychoice model the paper finds sufficient evi-dence to support the Self-Fulfilling Prophecynotion.
The effect of Emergency Liquidity Assistance (ELA) on bank lendingduring the euro area crisis
Working Paper Νο. 278Heather D. Gibson, Stephen G. Hall, Pavlos Petroulas, Vassilis Spiliotopoulos and George S. Tavlas
The paper examines the impact of emergencyliquidity assistance (ELA) on bank lending ineleven euro area countries during the finan-cial crisis. With the intensification of the cri-sis, ELA took on a pivotal role in some coun-tries. However, assessments of the quantita-tive impact of ELA in the literature are non-existent. The authors estimate a structuralpanel model for the determination of banklending, which includes the amount of ELA
received by each bank, allowing them toinvestigate the direct effect of ELA on lend-ing. The model corrects a misspecificationfound in the prototype model used in the lit-erature. The authors then undertake a VARanalysis, which allows them to address theeffect of ELA on GDP. Finally, they examinespillover effects among banks, indicating thatELA generated positive spillovers to otherbanks.
On the controversy over the origins of the Chicago Plan for 100 percent reserves
Working Paper Νο. 279George S. Tavlas
51Economic BulletinJuly 2020
Perceived vs actual financial crisis and bank credit standards: is there any indication of self-fulfilling prophecy?
Working Paper Νο. 277Dimitrios Anastasiou, Zacharias Bragoudakis and Stelios Giannoulakis
128
The idea of 100 percent reserve requirementsagainst demand deposits received a renewedimpetus following the 2007-08 financial crisis.
In 1933, a group of University of Chicago econ-omists, led by Frank Knight and Henry Simons,circulated two memoranda that called for 100
51Economic Bulletin
July 2020 129
A study of the effect of data transformation and “linearization” on time series forecasts.A practical approach
Working Paper Νο. 280Alexandros E. Milionis and Nikolaos G. Galanopoulos
percent reserve requirements. The ideabecame known as the Chicago Plan of BankingReform. That same idea had been proposed in1926 by Frederick Soddy, a Nobel Laureate inchemistry, in his book Wealth, Virtual Wealthand Debt. Soddy claimed precedence, a claimthat caught on. The paper provides evidence
showing that Knight, and probably Simons, con-ceived the idea of 100 percent reserves prior tothe publication of Soddy’s 1926 book. By 1934,however, Simons raised concerns that 100 per-cent reserves would not be sufficient in a worldwhere financial markets could innovatearound legal restrictions on banks.
Very often in actual macroeconomic timeseries there are causes that disrupt the under-lying stochastic process and their treatment isknown as “linearization”. In addition, variancenon-stationarity is in many cases also presentin such series and is removed by proper datatransformation. The impact of either (datatransformation – linearization) on the qualityof forecasts has not been adequately studied todate. This work examines their effect on uni-
variate forecasting considering each one sep-arately, as well as in combination, using twentyof the most important time series for the Greekeconomy. Empirical findings show a significantimprovement in forecasts’ confidence intervals,but no substantial improvement in point fore-casts. Furthermore, the combined transfor-mation-linearization procedure improves sub-stantially the non-normality problem encoun-tered in many macroeconomic time series.
Did the absence of a central bank backstop in the sovereign bond markets exacerbate spillovers during the euro area crisis?
Working Paper Νο. 281Heather D. Gibson, Stephen G. Hall, Deborah GeFang, Pavlos Petroulas and George S. Tavlas
The euro area sovereign debt crisis was char-acterised by feedback loops between (1) sov-ereign bond ratings and sovereign spreads insingle jurisdictions and (2) sovereign spreadsand ratings among jurisdictions. One expla-nation of this circumstance is that the ECB wasunable to perform the role of lender of lastresort in the sovereign bond markets duringthe crisis. The authors provide a spatial frame-
work that allows distinguishing among Euro-pean countries whose central banks were per-mitted to function as lender of last resort inthose markets and countries whose centralbanks were not permitted to do so. Theirresults are consistent with the view that theabsence of a central bank backstop in the sov-ereign bond markets exacerbated feedbackloops.
The authors provide a new way of deriving anumber of dynamic unobserved factors from a setof variables. They also show how standard prin-cipal components may be expressed in state spaceform and estimated using the Kalman filter. Toillustrate their procedure they perform two exer-cises. First, they use it to estimate a measure of
the current account imbalances among northernand southern euro area countries that developedduring the period leading up to the outbreak of the euro area crisis, before looking at adjust-ment in the post-crisis period. Second, they showhow these dynamic factors can improve fore-casting of the euro-dollar exchange rate.
51Economic BulletinJuly 2020130
A suggestion for a Dynamic Multi Factor Model (DMFM)
Working Paper Νο. 282Heather D. Gibson, Stephen G. Hall and George S. Tavlas
51Economic Bulletin
July 2020 131
Boutos Yannis, “Economic stabilisation and growth prospects”, No. 3, February 1994.Papademos Lucas, “Growth with stability: the role of monetary policy”, No. 5, March 1995.Voridis Hercules, “The special nature of banks and the transmission mechanism of monetary
policy: A review of recent literature”, No. 5, March 1995.Filippides Anastasios, Kyriakopoulos Panayotis and Moschos Dimitrios, “Bank of Greece mon-
etary policy instruments”, No. 6, November 1995.Haralabides Michael, Hardouvelis Gikas and Papageorgiou George, “Changeover to the single
currency: Prospects and challenges for credit institutions in Greece”, No. 6, November 1995.Karabalis Nikos, “Harmonisation of Consumer Price Indices in EU countries”, No. 7, March 1996.Public Sector Accounts Department, Studies, Planning and Operations Development Office,
“Government Securities Markets”, No. 7, March 1996.Saccomanni Fabrizio, “Opportunities and challenges in the process of European monetary inte-
gration”, No. 7, March 1996.Papademos Lucas, “Challenges facing monetary policy on the road to EMU”, No. 8, November
1996.Information Systems and Organisation Department, “Developments in EU Payment Systems:
TARGET and the Hellenic system Hermes”, No. 8, November 1996.Brissimis Sophocles and Gibson Heather, “Monetary policy, capital flows and Greek disinflation”,
No. 9, March 1997.Sabethai Isaac, “From contractual earnings to labour costs: incomes policy, collective bargain-
ing and inflation (1991-1996)”, No. 9, March 1997.Hall S.G. and Zonzilos Nicholas, “The output gap and inflation in Greece”, No. 9, March 1997.Pantelidis Evangelos, “The methodology of the new balance of payments compilation system of
the Bank of Greece”, No. 9, March 1997.Papademos Lucas, “The globalisation of financial markets and the conduct of economic and mon-
etary policy”, No. 10, December 1997.Brissimis Sophocles and Kastrisianakis Efstratios, “Is there a credit channel in the Greek econ-
omy?”, No. 10, December 1997.Gibson Heather, “The relationship between the financial system and the real economy”, No. 10,
December 1997.Hondroyiannis George and Papapetrou Evangelia, “The causal relationship between consumer
and wholesale prices in Greece”, No. 10, December 1997.Hardy Daniel and Simigiannis George, “Competition and efficiency of the Greek banking sys-
tem”, No. 11, June 1998.Brissimis Sophocles and Gibson Heather, “What can the yield curve tell us about inflation?”,
No. 11, June 1998.Pantazidis Stelios, “Inflation, investment and economic growth in Greece”, No. 11, June 1998.Mitrakos Theodoros, “The contribution of income sources to overall inequality”, No. 11, June 1998.Garganas Nicholas, “Greece and EMU: prospects and challenges”, No. 12, December 1998.Brissimis Sophocles, Sideris Dimitris and Voumvaki Fragiska, “Purchasing power parity as a long-
run relationship: an empirical investigation of the Greek case”, No. 12, December 1998.Manassaki Anna, “European Union transfers to Greece: historical background and prospects”,
No. 12, December 1998.Lazaretou Sophia, “The drachma on the metallic monetary standards: lessons from the past”,
No. 13, July 1999.Hondroyiannis George, “The causality between government spending and government revenue
in Greece”, No. 13, July 1999.Mitrakos Theodoros and Tsakloglou Panos, “The distributional impact of excise duties”, No. 13,
July 1999.
AR T I C L E S P UB L I S H ED I N P R E V I OU S I S S U E S O FTH E E CONOM I C BU L L E T I N
51Economic BulletinJuly 2020132
Dellas Harris, “ECB monetary policy strategy and macroeconomic volatility in Greece”, No. 13,July 1999.
Papazoglou Christos, “The real exchange rate and economic activity: is the hard-drachma pol-icy necessarily contractionary?”, No. 14, December 1999.
Gibson Heather and Lazaretou Sophia, “Leading inflation indicators for Greece: an evaluationof their predictive power”, No. 14, December 1999.
Karapappas Andreas and Milionis Alexandros, “Estimation and analysis of external debt in theprivate sector”, No. 14, December 1999.
Papademos Lucas, “From the drachma to the euro”, No. 15, July 2000.Zonzilos Nicholas, “The Greek Phillips curve and alternative measures of the NAIRU”, No. 15,
July 2000.Pantazidis Stelios, “The sustainability of the current account deficit”, No. 15, July 2000.Hondroyiannis George, “Investigating causality between prices and wages in Greece”, No. 15, July 2000.Sabethai Isaac, “The Greek labour market: features, problems and policies (with emphasis on
labour market flexibility and on combatting unemployment in the context of non-inflationarygrowth)”, No. 16, December 2000.
Brissimi Dimitra and Brissimis Sophocles, “The problem of unemployment in the EuropeanUnion”, No. 16, December 2000.
Bardakas Joanna, “Financial deregulation, liquidity constraints and private consumption inGreece”, No. 16, December 2000.
Lazaretou Sophia and Brissimis Sophocles, “Fiscal rules and stabilisation policy in the euro area”,No. 17, July 2001.
Lolos Sarantis-Evangelos, “The role of European structural resources in the development of theGreek economy”, No. 17, July 2001.
Papazoglou Christos, “Regional economic integration and inflows of foreign direct investment:the European experience”, No. 17, July 2001.
Garganas Nicholas, “The future of Europe and EMU”, No. 18, December 2001.Brissimis Sophocles and Papadopoulou Dafni-Marina, “The physical introduction of euro ban-
knotes and coins in Greece”, No. 18, December 2001.Zonzilos Nicholas, “The monetary transmission mechanism – the ECB experiment: results for
Greece”, No. 18, December 2001.Brissimis Sophocles, Kamberoglou Nicos and Simigiannis George, “Bank-specific characteris-
tics and their relevance to monetary policy transmission”, No. 18, December 2001.Tsaveas Nicholas, “Exchange rate regimes and exchange rate policy in South Eastern Europe”,
No. 18, December 2001.Demenagas Nicholas and Gibson Heather, “Competition in Greek banking: an empirical study
for the period 1993-1999”, No. 19, July 2002.Hondroyiannis George, “Economic and demographic determinants of private savings in Greece”,
No. 19, July 2002.Gatzonas Efthymios and Nonika Kalliopi, “Eligible assets and management of collateral in the
context of central bank monetary policy operations”, No. 19, July 2002.Voridis Hercules, Angelopoulou Eleni and Skotida Ifigeneia, “Monetary policy in Greece 1990-
2000 through the publications of the Bank of Greece”, No. 20, January 2003.Kaplanoglou Georgia and Newbery David, “The distributional impact of indirect taxation in
Greece”, No. 21, July 2003.Papapetrou Evangelia, “Wage differentials between the public and the private sector in Greece”,
No. 21, July 2003.Athanasoglou Panayiotis and Brissimis Sophocles, “The effect of mergers and acquisitions on
bank efficiency in Greece”, No. 22, January 2004.
Lazaretou Sophia, “Monetary system and macroeconomic policy in Greece: 1833-2003”, No. 22,January 2004.
Karabalis Nikos and Kondelis Euripides, “Alternative measures of inflation”, No. 22, January 2004.Papaspyrou Theodoros, “EMU strategies for new Member States: the role of Exchange Rate
Mechanism II”, No. 23, July 2004.Mitrakos Theodoros, “Education and economic inequalities”, No. 23, July 2004.Papapetrou Evangelia, “Gender wage differentials in Greece”, No. 23, July 2004.Gibson Heather, “Greek banking profitability: recent developments”, No. 24, January 2005.Athanasoglou Panayiotis, Asimakopoulos Ioannis and Georgiou Evangelia, “The effect of merger
and acquisition announcement on Greek bank stock returns”, No. 24, January 2005.Mitrakos Theodoros and Zografakis Stavros, “The redistributional impact of inflation in Greece”,
No. 24, January 2005.Theodossiou Ioannis and Pouliakas Konstantinos, “Socio-economic differences in the job sat-
isfaction of high-paid and low-paid workers in Greece”, No. 24, January 2005.Garganas Nicholas, “Adjusting to the single monetary policy of the ECB”, No. 25, August 2005.Mitrakos Theodoros, Simigiannis Georgios and Tzamourani Panagiota, “Indebtedness of Greek
households: evidence from a survey”, No. 25, August 2005.Nicolitsas Daphne, “Per capita income, productivity and labour market participation: recent devel-
opments in Greece”, No. 25, August 2005.Nicolitsas Daphne, “Female labour force participation in Greece: developments and determin-
ing factors”, No. 26, January 2006.Mitrakos Theodoros and Zonzilos Nicholas, “The impact of exogenous shocks on the dynamics and
persistence of inflation: a macroeconomic model-based approach for Greece”, No. 26, January 2006.Athanasoglou Panayiotis, Asimakopoulos Ioannis and Siriopoulos Konstantinos, “External financ-
ing, growth and capital structure of the firms listed on the Athens Exchange”, No. 26, January 2006.Mitrakos Theodoros and Nicolitsas Daphne, “Long-term unemployment in Greece: developments,
incidence and composition”, No. 27, July 2006.Kalfaoglou Faidon, “Stress testing of the Greek banking system”, No. 27, July 2006.Pantelidis Evangelos and Kouvatseas Georgios, “Frontier survey on travel expenditure: method-
ology, presentation and output assessment (2003-2005)”, No. 27, July 2006.Brissimis Sophocles and Vlassopoulos Thomas, “Determinants of bank interest rates and com-
parisons between Greece and the euro area”, No. 28, February 2007.Simigiannis George and Tzamourani Panagiota, “Borrowing and socio-economic characteristics of
households: results of sample surveys carried out by the Bank of Greece”, No. 28, February 2007.Papapetrou Evangelia, “Education, labour market and wage differentials in Greece”, No. 28, Feb-
ruary 2007.Christodoulakis George, “The evolution of credit risk: phenomena, methods and management”,
No. 28, February 2007.Karabalis Nikos and Kondelis Euripides, “Inflation measurement in Greece”, No. 28, February 2007.Kapopoulos Panayotis and Lazaretou Sophia, “Foreign bank presence: the experience of South-
East European countries during the transition process”, No. 29, October 2007.Nicolitsas Daphne, “Youth participation in the Greek labour market: developments and obsta-
cles”, No. 29, October 2007.Albani Maria, Zonzilos Nicholas and Bragoudakis Zacharias, “An operational framework for the
short-term forecasting of inflation”, No. 29, October 2007.Asimakopoulos Ioannis, Brissimis Sophocles and Delis Manthos, “The efficiency of the Greek
banking system and its determinants”, No. 30, May 2008.Balfoussia Hiona, “Stock market integration: the Athens Exchange in the European financial
market”, No. 30, May 2008.
51Economic Bulletin
July 2020 133
Mitrakos Theodoros, “Child poverty: recent developments and determinants”, No. 30, May 2008.Asimakopoulos Ioannis, Lalountas Dionysis and Siriopoulos Constantinos, “The determinants
of firm survival in the Athens Exchange”, No. 31, November 2008.Petroulas Pavlos, “Foreign direct investment in Greece: Productivity and Technology diffusion”,
No. 31, November 2008.Anastasatos Tasos and Manou Constantina, “Speculative attacks on the drachma and the
changeover to the euro”, No. 31, November 2008.Mitrakos Theodoros and Simigiannis George, “The determinants of Greek household indebt-
edness and financial stress”, No. 32, May 2009.Papazoglou Christos, “Is Greece’s export performance really low?”, No. 32, May 2009.Zonzilos Nicholas, Bragoudakis Zacharias and Pavlou Georgia, “An analysis of the revisions of
first (flash) quarterly national account data releases for Greece”, No. 32, May 2009.Rapanos Vassilis and Kaplanoglou Georgia, “Independent fiscal councils and their possible role
in Greece”, No. 33, May 2010.Mitrakos Theodoros, Tsakloglou Panos and Cholezas Ioannis, “Determinants of youth unem-
ployment in Greece with an emphasis on tertiary education graduates”, No. 33, May 2010.Gibson Heather and Balfoussia Hiona, “Inflation and nominal uncertainty: the case of Greece”,
No. 33, May 2010.Migiakis Petros, “Determinants of the Greek stock-bond correlation”, No. 33, May 2010.Mitrakos Theodoros, Tsakloglou Panos and Cholezas Ioannis, “Determinants of the wage rates in
Greece with an emphasis on the wages of tertiary education graduates”, No. 34, September 2010.Bragoudakis Zacharias and Panagiotou Stelios, “Determinants of the receipts from shipping
services: the case of Greece”, No. 34, September 2010.Gibson Heather and Balfousia Hiona, “The impact of nominal and real uncertainty on
macroeconomic aggregates in Greece”, No. 34, September 2010.Manesiotis Vasilios, “Numerical fiscal rules in practice”, No. 35, June 2011.Vasardani Melina, “Tax evasion in Greece: an overview”, No. 35, June 2011.Kalfaoglou Faidon, “The usefulness of stress testing exercises for assessing the soundness of the
banking system”, No. 35, June 2011.Nicolitsas Daphne, “On-the-job training in Greece: a brief overview”, No. 35, June 2011.Papaspyrou Theodoros, “EMU sustainability and the prospects of peripheral economies”, No.
36, April 2012.Kanellopoulos Costas Ν., “Employment and worker flows during the financial crisis”, No. 36,
April 2012.Kalfaoglou Faidon, “Bank capital adequacy framework”, No. 36, April 2012.Bragoudakis Zacharias and Sideris Dimitris, “Do retail gasoline prices adjust symmetrically to
crude oil price changes? The case of the Greek oil market”, No. 37, December 2012.Kanellopoulos Costas Ν., “The size and structure of uninsured employment”, No. 37, December 2012.Migiakis Petros M., “Reviewing the proposals for common bond issuances by the euro area sov-
ereign states under a long-term perspective”, No. 37, December 2012.Karabalis Nikos and Kondelis Euripides, “Indirect tax increases and their impact on inflation
over 2010-2012”, No. 38, November 2013.Vourvachaki Evangelia, “Structural reform of the road haulage sector”, No. 38, November 2013. Kanellopoulos Nikolaos C., Mitrakos Theodoros, Tsakloglou Panos and Cholezas Ioannis, “The
impact of the current crisis on private returns to education in Greece”, No. 38, November 2013. Papapetrou Evangelia and Bakas Dimitrios, “The Greek labour market during the crisis: Unem-
ployment, employment and labour force participation”, No. 38, November 2013. Tzamourani Panagiota, “The Household Finance Survey: Description of the 2009 survey and main
results on households’ income, wealth and debt in Greece”, No. 38, November 2013.
51Economic BulletinJuly 2020134
Dimitropoulou Dimitra, Koutsomanoli-Filippaki Anastasia, Charalambakis Evangelos and Agge-lis Georgios, “The financing of Greek enterprises before and during the crisis”, No. 39, July 2014.
Kanellopoulos Costas, “The Greek private sector labour market during the crisis”, No. 39, July 2014. Kalfaoglou Faidon, “European Banking Union: “Europeanising” banks’ financial safety net”, No.
39, July 2014. Lazaretou Sofia, “The smart economy: cultural and creative industries in Greece. Can they be
a way out of the crisis?”, No. 39, July 2014. Chaireti Aikaterini, “The housing situation in Greece during the crisis years 2008-2012”, No. 39,
July 2014. Tagkalakis Athanasios, “Tax buoyancy in Greece”, No. 40, December 2014. Bragoudakis G. Zaharias, “An empirical investigation of the relationship between unit labour
cost and price developments: The case of Greece”, No. 40, December 2014. Migiakis M. Petros, “The international financial markets as a source of funding for Greek non-
financial corporations”, No. 40, December 2014. Kalyvitis Sarantis, “Estimating the quality of Greek exports: A non-price based approach”, No.
40, December 2014. Kanellopoulos Costas, “The effects of minimum wages on wages and employment”, No. 41, July
2015. Georgiou Evangelia, “The use of cash and electronic payment instruments in the economy”, No.
41, July 2015. Kalfaoglou Faidon, “Alternative NPLs resolution regimes. A case studies approach”, No. 41, July
2015. Lazaretou Sophia, “How can the use of historical macroeconomic data series support economic
analysis and policy making? The new macro history database 1833-1949 as part of the large data-base of Southeast European Countries”, No. 41, July 2015.
Kasimati Evangelia and Sideris Dimitrios, “Towards a new growth model for tourism: structuralreforms and the tourism product in Greece during the crisis (2008-2014)”, No. 42, December 2015.
Tagkalakis Athanasios, “The determinants of unemployment dynamics in Greece”, No. 42,December 2015.
Chaireti Aikaterini, “The Greek primary sector from 1995 to 2014”, No. 42, December 2015.]Vettas Nikos, Giotopoulos Ioannis, Valavanioti Evangelia and Danchev Svetoslav, “The deter-
minants of new firms’ export performance”, No. 43, July 2016.Belli Styliani and Backinezos Constantina, “The transition to the new methodology for the com-
pilation of balance of payments statistics – BPM6”, No. 43, July 2016.Lazaretou Sophia, “The Greek brain drain: the new pattern of Greek emigration during the recent
crisis”, No. 43, July 2016.Kalfaoglou Faidon, “Bank recapitalisation: a necessary but not sufficient condition for resum-
ing lending”, No. 43, July 2016.Manou Konstantina and Papapetrou Evangelia, “The economic behaviour of households in
Greece: recent developments and prospects”, No. 44, December 2016.Papazoglou Christos, Sideris Dimitrios and Tsagklas Orestis, “Europe as an Optimum Currency
Area: the experience of the Baltic countries”, No. 44, December 2016.Anyfantaki Sofia, “The evolution of Financial Technology (FinTech)”, No. 44, December 2016.Chatzivasiloglou Ioannis, “Introduction to Solvency II for (re)insurance undertakings”, No. 44,
December 2016.Gibson Heather D. and Pavlou Georgia, “Exporting and performance: evidence from Greek
firms”, No. 45, July 2017.Charalambakis Evangelos, “How did the Greek financial crisis impact on households? A com-
parison between the two waves of the HFCS”, No. 45, July 2017.
51Economic Bulletin
July 2020 135
Kosma Theodora, Papapetrou Evangelia, Pavlou Georgia, Tsochatzi Christina and ZioutouPinelopi, “Labour market adjustments and reforms in Greece during the crisis: findings fromthe third wave of the Wage Dynamics Survey”, No. 45, July 2017.
Chatzivasiloglou Ioannis, “The valuation of assets and liabilities of (re)insurance undertakingsunder Solvency II”, No. 45, July 2017.
Louzis Dimitrios, “The macroeconomic effects of unconventional monetary policy in the euroarea using non-linear models”, No. 46, December 2017.
Albani Maria and Anyfantaki Sofia, “Research and innovation in Greece”, No. 46, December 2017.Vettas Nikos, Vassiliadis Michael, Papadakis Manos and Peppas Konstantinos, “The impact of
reforms on the sectors of tradable goods and services”, No. 46, December 2017.Chaireti Aikaterini, “Regional household accounts for the years 2000-2014”, No. 46, December
2017.Balfoussia Hiona and Papageorgiou Dimitris, “Real and financial cycles in the Greek economy”,
No. 47, July 2018.Vasardani Melina and Dimitropoulou Dimitra, “The utilisation of EU structural funds in Greece”,
No. 47, July 2018.Zioutou Pinelopi and Sideris Dimitrios, “The energy sector: developments and prospects”, No.
47, July 2018.Bragoudakis G. Zaharias, “Are price adjustments asymmetric in basic food categories? The case
of the Greek food market”, No. 47, July 2018. Georgiou Evangelia, “Financing constraints and firm characteristics: evidence from Greek SMEs”,
No. 48, December 2018.Dellis Konstantinos, “Structural determinants of FDI inflows and the case of Greece”, No. 48,
December 2018.Louzis Dimitrios P., “Greek GDP revisions and short-term forecasting”, No. 48, December 2018.Kosma Theodora, Petroulas Pavlos and Vourvachaki Evangelia, “Job flows in Greece during the
recent years”, No. 49, July 2019.Papaspyrou Marios and Petralias Athanasios, “The Greek shipping estimation model”, No. 49,
July 2019.Mpardaka Ioanna and Papazoglou Christos, “The determinants of Greece’s export supply of oil”,
No. 49, July 2019.Kasimati Evangelia and Veraros Nikolaos, “Raising capital through issuance of common shares
by Greek-controlled maritime companies in US capital markets”, No. 49, July 2019.Albani Maria, Anyfantaki Sofia and Lazaretou Sophia, “How do digital technologies drive
Greece’s economic growth? Opportunities and challenges”, No. 49, July 2019.Gibson Heather D., Pavlou Georgia, Tsochatzi Christina and Vasardani Melina, “Greece’s inte-
gration into global value chains”, No. 50, December 2019.Backinezos Constantina, Panagiotou Stelios and Rentifi Athina, “Greek export performance: a
constant market share analysis”, No. 50, December 2019.Vettas Nikos, Valavanioti Evangelia, Peppas Konstantinos and Vasileiadis Michail, “The evo-
lution of new firms’ characteristics in Greece before and during the Economic Adjustment Pro-grammes”, No. 50, December 2019.
Kalfaoglou Faidon, “Cryptoassets: potential implications for financial stability”, No. 50, Decem-ber 2019.
51Economic BulletinJuly 2020136
BA
NK
OF
GR
EEC
E
EUR
OSY
STE
M
ISSN: 2654 - 1904 (online)
JULY
20
20
EC
ON
OM
IC B
UL
LE
TIN
No
51
ISSN: 1105 - 9729 (print)